Abstract:Objective To explore the risk factors for pocket infection in patients with permanent pacemaker implantation (PPI), and construct a personalized risk prediction model. Methods Retrospective cohort study method was adopted to analyze clinical data and follow-up data of patients who received PPI in the department of cardiovascular diseases of a tertiary first-class hospital in Anhui Province from January 2016 to May 2018. According to the occurrence of pocket infection one year after operation, patients were divided into infection group and non-infection group. Independent risk factors for post-operative pocket infection in patients with PPI were analyzed with lasso regression and logistic regression analysis. Results A total of 322 patients with PPI were included in analysis, incidence of post-operative pocket infection was 7.45% (n=24). Lasso regression and multivariate logistic regression model analysis results showed that age, duration of operation, operation times and poor wound healing were independent risk factors for post-operative pocket infection (all P < 0.05). The constructed nomogram model was internally verified by Bootstrp self-sampling for 1 000 times, after internal verification, C-statistic of nomogram model was 0.869 (95%CI: 0.773-0.947), suggesting that the nomogram has good discrimination. Calibration curve analysis showed that the average absolute error between prediction probability of nomogram model and the actual occurrence of PPI in patients was 0.018, which indicated that nomogram has a good calibration degree. Decision curve analysis (DCA) showed that when probability threshold of nomogram model was 0.03-0.62, clinical net benefit rate of patients was the highest, which indicated that nomogram model had better clinical applicability. Conclusion The occurrence of post-operative pocket infection in patients with PPI is affected by multiple risk factors, nomogram can integrate the risk factors for post-operative pocket infection and be used to personalized prediction of risk of post-operative pocket infection in patients, which has high clinical application value.