Epidemiology of Central Venous Catheter Utilization: The Devil is in the Details

Background: Aggregate number of device days is an appropriate denominator for surveillance reports of central line-associated bloodstream infections (CLABSIs). However, the full risk milieu of central venous catheter (CVC) complications encompasses many other details of CVC utilization.

Objective: To conduct a detailed epidemiologic survey of CVC utilization in order to more fully understand how to prevent complications.

Methods: In 2008, using a semi-random sample, we prospectively collected the following data on all non-tunneled CVC insertions for 6 months in the medical/surgical unit of a Midwestern, academic, level 1 trauma center: patient demographics; CVC type, site, and side; skin prep and dressing type; insertion and removal dates; documentation of procedure note and CVC bundle checklist; major complications including infectious, mechanical, and thrombotic.

Results: During the study period, 203 CVCs were placed in 144 patients. The mean (SD) age of the patients was 54 (15) years with 49% being non-white and 51% female. Of the 203 CVCs inserted, 78% were traditional CVCs, 14% were temporary hemodialysis catheters, and 8% peripherally inserted central catheters (PICCs). The site of insertion was 50% subclavian vein (SV), 32% internal jugular (IJ) vein, 9% femoral vein (FV), and 8% PICC. Fifty-eight percent were inserted on the right side versus the left. Seventy-eight percent were triple lumen CVCs and 22% dual lumen. A procedure note was documented for 73% of insertions, whereas the CVC bundle checklist was documented for only 13% of insertions. On a per line basis, the mean (SD) time that a CVC was left in place was 11.1 (7.5) days with a range of 43 days. On a per patient basis, the mean (SD) time that a patient had one or more CVCs in place was 15.7 (14.3) days with a range of 67 days. The length of time a CVC was left in place varied by site with FV being the shortest (mean 5.9 days, range 12 days) and IJ the longest (mean 13.6 days, range 43 days). Major complications included 26 CLABSIs, 3 pneumothoraces, 1 hemorrhage, and 1 thrombosis. In multivariate analysis, longer length of catheterization (OR 1.4 per 7 days, P=0.05) and placement on left side of body (OR 2.8, P=0.02) were independent predictors of CLABSI.

Conclusions: How long an individual CVC is left in place remains an independent predictor of CLABSI. This important piece of data is not reflected in aggregate device days. We also found that CVC placement on the left side of the body was in independent predictor of CLABSI. This novel finding merits further study. We hypothesize that it could be due to slight technical difficulties of a right-handed person putting a CVC in on the left side of a patient. It could also be a marker of prior CVC insertion on the right. As we move toward more robust electronic medical records, tracking details of individual CVC characteristics in real time could help mitigate the risk of infectious complications.