Insights and Characteristics of Pediatric Central Venous Catheter Malfunction and Catheter-Associated Deep Vein Thrombosis: A Subanalysis of the Circle Study

Introduction: Central venous catheters (CVCs) are increasingly placed in children for administration of medications, fluids, parenteral nutrition as well as frequent blood collection. CVC type or insertion technique, underlying medical conditions and type of infused medications can impact CVC function and longevity. CVC malfunction is known to be associated with deep vein thrombosis (DVT). Our previous study of children with CVCs (Jaffray, et al, Blood 2020), found those with a CVC malfunction were six times more likely to develop a DVT. The aim of this subanalysis was to evaluate risk factors for CVC malfunction and subsequently the development of a DVT in children with peripherally inserted central catheters (PICCs) and tunneled lines (TLs) within our Clot Incidence Rates in Central Lines (CIRCLE) study.

Methods: This is a subanalysis of the multicenter, prospective CIRCLE study conducted from October 2013 to June 208. Randomly selected children aged 6 months to <18 years of age with newly placed PICCs or TLs were enrolled and monitored until CVC removal or end of the study period. A CVC malfunction was defined as any CVC that required tissue plasminogen activator (t-PA) or had malfunction reported as ‘blockage’ or ‘mechanical’. CVC related risk factors for DVT included type, size, material, length, number of lumens of CVC and insertion technique including number of insertion attempts, placement via ultrasound guidance and whether CVC was placed by an interventional radiology (IR) technique. Patient related factors included age, medical history, vein accessed, location of CVC tip, having a catheter associated blood stream infection (CLABSI), receiving total parental nutrition or anticoagulation (other than routine heparin flushes). DVT was confirmed on radiologic examination of those subjects with symptoms of CVC-associated DVT. To examine associations between risk factors and CVC malfunction, we performed survival analysis by fitting univariate and multivariate Weibull regression models. From these models, we obtained hazard ratios (HR) and their respective 95% confidence intervals (CI).

Results: A total of 1951 CVCs were included in this analysis. There were 380 (19.5%) CVCs with at least one malfunction. Median [interquartile range (IQR)] time from CVC insertion to the first malfunction was 20 [7.75-57] days. The majority of CVCs with malfunction had one episode (n= 219, 57.6%), followed by two malfunctions (n= 75, 19.7%) and 22.7%(n=86) had three or more malfunctions. There was a significantly higher rate of malfunction in PICCs as compared to TLs (HR=2.14, 95% CI:1.71-2.68, p value<0.0001). On multivariable analysis, risk factors associated with an increased risk of malfunction included previous CVC (HR=1.5, 95% CI:1.2-1.9, p=0.002), CVC size < 4 French (HR=3.23, 95% CI:2.51-4.17, p<0.0001), having two or more catheter lumens (HR=2.9, 95% CI:2.3-3.6, p<0.0001) and more than two insertion attempts (HR=2.2, 95% CI:1.4-3.5, p=0.0005). CVCs placed by IR (HR=0.4, 95% CI:0.3-0.5, p<0.0001) and the presence of a CLABSI (HR=0.6, 95% CI:0.4-0.9, p=0.009) were associated with significantly lower risk of CVC malfunction.

DVT was diagnosed in 22 (5.8%) CVCs with a malfunction and the median (IQR) time to diagnosis of DVT after CVC malfunction was 6 (1-18.5) days. The HR of DVT development for those requiring one t-PA dose versus those receiving no t-PA was 0.87 (95% CI:0.16-4.57, p= 0.87), compared to a HR of 1.23 (95% CI:0.21-7.23, p= 0.81) for two doses, 1.15 for three doses (95% CI:0.15-8.68, p=0.89) and 3.37 for four or more doses (95% CI:0.66-17.27, p=0.11) of t-PA.

Conclusion: Our previous study identified that CVCs with malfunction were more likely to develop a DVT. PICCs had nearly a 9-fold increased risk of DVT as compared to TLs. In this sub-analysis, nearly one fifth of CVCs placed in children had a malfunction. Malfunction risk factors included a previous CVC, smaller lumen size, multi-lumen CVCs, multiple insertion attempts and CVCs not being inserted by IR technique. Interestingly, the presence of a CLABSI was associated with lower rates of CVC malfunction which could be attributed to frequent CVC access that maintains patency or use of medication locks for CLABSI. Although not statistically significant, likely due to small numbers, CVC-associated DVT seems to increase with multiple doses of t-PA which should raise suspicion to screen for DVT if a CVC requires repeated t-PA instillation.