Risk factors analysis of central venous catheter-related thrombosis in critically ill patients and development of nomogram prediction model / 中华危重病急救医学

ABSTRACT

Objective:To analyze the risk factors of central venous catheter-related thrombosis (CRT) in critically ill patients and develop the model of a nomogram.Methods:A prospective investigation study was conducted on 385 critically ill patients who received central venous catheters during hospitalization in Hengshui People's Hospital from May 2018 to March 2021. Color Doppler ultrasonography was performed daily after catheterization. Patients were divided into thrombosis group and non-thrombosis group according to whether CRT was formed. The patient's gender, age, body mass index (BMI), acute physiology and chronic health evaluationⅡ(APACHEⅡ) score, complications, existing tumor, D-dimer level on the 3rd day after catheterization, maximum velocity of right internal jugular vein on the 3rd day after catheterization, mechanical ventilation time, and catheter indwelling time were recorded, and the differences of above indexes between the two groups were compared. Multivariate Logistic regression was performed on the influencing factors with statistical differences between the two groups to establish the nomogram prediction. The receiver operator characteristic curve (ROC curve) and calibration curve were used to evaluate the predictive power of the model.Results:The incidence of central venous CRT in critically ill patients was 16.1% (62/385). Compared with non-thrombosis patients, the thrombosis group patients had higher APACHEⅡscore, the proportion of existing tumor, and D-dimer level on the 3rd day after catheterization [APACHEⅡscore 17 (15, 19) vs. 15 (12, 18), the proportion of existing tumor 51.6% (32/62) vs. 35.3% (114/323), D-dimer (mg/L) 0.84 (0.64, 0.94) vs. 0.57 (0.44, 0.76), all P < 0.05], the maximum flow rate of right internal jugular vein was slower on the 3rd day after catheterization [cm/s 14 (13, 15) vs. 16 (14, 18), P < 0.05]. Univariate analysis showed that high APACHEⅡscore, critical patients with existing tumor, high D-dimer level on the 3rd day after catheterization, and slow maximum flow rate of right internal jugular vein on the 3rd day after catheterization were more likely to develop central venous CRT. Further multivariate Logistic regression analysis showed that high APACHEⅡscore, existing tumor, high D-dimer level on the 3rd day after catheterization and slow maximum flow rate of right internal jugular vein on the 3rd day after catheterization were independent risk factors for central venous CRT in critical patients [odds ratio ( OR) and 95% confidence interval (95% CI) were 0.876 (0.801-0.957), 0.482 (0.259-0.895), 0.039 (0.011-0.139), 1.401 (1.218-1.611), and P values were 0.003, 0.021, < 0.001, < 0.001, respectively]. According to the results of multivariate analysis, the prediction model of the nomogram was constructed. The area under ROC curve (AUC) was 0.820, 95% CI was 0.767-0.872, P < 0.001. The calibration curve showed that the prediction probability of central venous CRT nomogram model in critically ill patients had good consistency with the actual occurrence probability. Conclusions:Existing tumor, high APACHEⅡscore, elevated D-dimer on the 3rd day after catheterization, and decreased maximum velocity of right internal jugular vein on the 3rd day after catheterization are independent risk factors for central venous CRT in critical patients. The prediction model based on the proposed model has good clinical efficacy.