Study Of Influence Factors Of Peripherally Inserted Central Catheters Related Thrombosis And Preliminary Build Of Its Risk Assessment Model

Objective1. To understand the incidence rate of peripherally inserted central catheters related thrombosis.2. To analysis the risk factors of peripherally inserted central catheters related thrombosis.3. To build risk assessment model of peripherally inserted central catheters related thrombosis.4. To validate the predictive power of risk assessment model of peripherally inserted central catheters related thrombosis.Methods1. Screened the patient which inserted PICC in special clinic of the three grade general hospital in Fujian province within January 2012 to December 2014,retrospectively investigate the patients which was confirmed with PICC related thrombosis, analysis the incidence rate and clinical characteristics of PICC related thrombosis.2. We conducted a retrospective case control study, screened the patients with PICC related thrombosis which met the inclusive and exclusion criteria as case group,according to age and sex at the ratio of 1:2 screened the patient with PICC but not developed PICC related thrombosis in the same period as control group. Randomly selected 70 percent of case group(287) and matched control group(574) into model group, the rest 30 percent of case group(121) and matched control group(242) into validation group. We screened the independent risk factors of PICC related thrombosis by univariate and multivariate analysis.3. Follow the independent risk factors of PICC related thrombosis, assigned point values to each independent risk factor according to the odds ratio, build the risk assessment model of PICC related thrombosis. Using receiver operator characteristic curve and Hosmer-Lemeshow goodness-of-fit to validated the risk assessment in the model group and validation group.Results1. The incidence of PICC related thrombosis were 11.04%. The clinical characteristics of PICC related thrombosis as follows: male patients(59.28%) are more than female patients(40.72%); ages mainly range from 40 to 65; the digestive system tumor suffered the largest portion(35.5%), next was hematological tumor(31.4%) and breast tumor(15.2%).2. Univariate analysis demonstrated that deep vein thrombosis, prophylactic anticoagulation by heparin, indwelling time of catheter, chemotherapy, transfusion red blood cells, administration of haemostatic, and the counts of red blood cell,hemoglobin, cholesterol, albumin were associated with PICC related thrombosis(P<0.05).3. Multivariable analysis showed that prophylactic anticoagulation by heparin(OR=4.252, P=0.000, 95%CI=2.414~7.490), short indwelling time of catheter(OR=2.267, P=0.000, 95%CI=1.717~2.992), number of chemotherapy(OR=1.285,P=0.021, 95%CI=1.039~1.589), administration of haemostatic(OR=3.236, P=0.000,95%CI=1.732~6.046), ALB<35g/L(OR=1.697, P=0.012, 95%CI=1.126~2.557) were independent risk factors of PICC related thrombosis, APTT higher than normal was protect factors of PICC related thrombosis(OR=0.353, P=0.001,95%CI=0.189~0.659).4. According to odd ratio assigned values to each risk factors of PICC related thrombosis, build risk assessment model of PICC related thrombosis, risk score assigned as follow: About chemotherapy, when the number is 1 to 3, scored 1 point,when the number is 4 to 6, scored 2 points, when the number is 7 to 9, scored 3 points, when the number is over 9, scored 4 points. Second, about the indwell time of catheter,when indwell time is less than a 7 days, scored 8 points, when indwell time between 7to 30 days, scored 6 points, when indwell time between 31 to 90 days, scored 4 point,when indwell time more than 90 days, scored 2 point. Prophylactic anticoagulation by heparin scored 4 points, administration of haemostatic scored 3 points, ALB<35g/L scored 2 points.5. Using receiver operator characteristic curve to validate the risk assessment model,in model group, the area under curve(AUC) of the PICC related thrombosis risk assessment model was 0.695 by ROC, the cut off value was 4.5, the sensitivity and specificity was 54.96%, 75.90%, respectively(P<0.001). In validation group, AUC was 0.699, the sensitivity and specificity was 55.45%, 74.03%, respectively(P<0.001).6. Using Hosmer-Lemeshow goodness-of-fit to validate the risk assessment model, in model group, we found that the expected value and observed value of PICC related thrombosis had no significant difference(x2=3.291,P=0.857),the accuracy of prediction was 72.1%. In validation group, we found that the expected value and observed value of PICC related thrombosis had no significant difference(x2=9.115, P=0.333), the accuracy of prediction was 75.1%.Conclusions1. The incidence of PICC related thrombosis of the three grade general hospital in Fujian province was 11.04%.The clinical characteristics of PICC related thrombosis in the hospital as follow:ages mainly range from 40 to 65, most patients were female,the digestive system tumor suffered the largest portion, next was hematological tumor and breast tumor,hope it can provide preference for the clinics.2. The independent risk factors of PICC related thrombosis were prophylactic anticoagulation by heparin, short indwelling time of catheter, number of chemotherapy, administration of haemostatic, ALB<35g/L. And APTT higher than normal was protect factors of PICC related thrombosis.3. We build the risk assessment model of PICC related thrombosis preliminary, it can be used to evaluate the risk of PICC related thrombosis in patients with PICC, but the sensitivity of model is a little low, so it need to be improved in future study.