Abstract
Background: Central venous access devices (CVADs) are essential in the modern management of patients with haematological malignancies and solid tumors. An often serious complication is catheter-associated bloodstream infection (CA-BSI) which is a major cause of morbidity and mortality. We have previously demonstrated (JHI 2011;78:26) in a retrospective study of 1127 CVADs that insertion of lines on the right side of the body was associated with increased risk of CA-BSI (HR 1.60, 95%CI 1.05-2.44). We hypothesized that this finding related to the predominance of right-handedness in the patient (as for the general) population, which may result in greater movement and thus associated bacterial contamination of the catheterised limb.
Methods: We performed a prospective randomised controlled, non-blinded study at two centres to determine whether the side of CVAD insertion influences the incidence of CA-BSI with consenting patients randomly allocated to either dominant or non-dominant-side CVAD insertion. Any in- or out-patient >14yrs with cancer requiring a CVAD was eligible. Patients had to have both sides of the body available for CVAD insertion so patients with contraindications to CVAD insertion on one particular side of the body were excluded. The primary end-point of the study was the number of line days until CA-BSI, determined in a blinded fashion by two specialists according to standard Australia Council of Healthcare Standards definitions. Secondary end-points included catheter-related blood stream infection (CR-BSI, a stricter definition of line-related infection requiring a positive line tip culture or differential time to positivity of ≥ 2hrs between line and peripheral cultures) and CA-BSI restricted to patients with peripherally inserted central catheters (PICC lines). Analysis was by intent-to-treat.
Results: 640 CVADs were randomised to dominant (n=322) or non-dominant (n=318) side of insertion. Five subjects were subsequently excluded (dominant side n=2, non-dominant side n=3) leaving 635 for final analysis. 34% were female, median age was 56yrs and 9.3% were left handed. 60% had hematological malignancies and 40% had solid tumors. The CVAD type was a PICC line (67%), tunnelled Hickman line (23%) and non-tunnelled line (10%); 85% were double and 13% triple lumen lines. Baseline characteristics were all well balanced between randomization arms.
22% of CVADs were complicated by CA-BSI. The rate of CA-BSI per 1000 line days was 3.49 vs 3.66 in the non-dominant vs dominant group (HR 0.91; 95%CI, 0.65-1.28; p=0.602; Fig 1A). By multivariate analysis, the rate of CA-BSI was increased by: use of tunnelled Hickman lines compared to PICC lines (HR 2.05, 95%CI 1.45-2.91); having a haematological malignancy needing intensive chemotherapy or transplantation compared to non-GI solid tumors (HR 5.56; 95%CI 2.47-12.5); but not non-dominant compared to dominant side of line insertion (HR 0.94, 95%CI 0.69-1.36).
2.5% of CVADs were complicated by CR-BSI. The rate of CR-BSI per 1000 line days was 0.26 vs 0.55 in the non-dominant vs dominant group (HR 0.46; 95%CI, 0.15-1.34; p=0.153; Fig 1B). By multivariate analysis, the rate of CR-BSI was increased by use of tunnelled Hickman lines compared to PICC lines (HR 5.39, 95%CI 1.78-16.33) and there was a trend to increased CR-BSI in patients with the CVAD inserted in the dominant side (HR 0.37, 95%CI 0.12-1.15, p=0.084). There was no impact of patient diagnosis on CR-BSI. In analysis confined to PICC lines, there was no impact of side of insertion on CA-BSI or CR-BSI.
Because the trial design also allowed a randomized comparison between right and left side of insertion (as opposed to non-dominant vs dominant side) we examined if a right sided CVAD affected outcomes (Fig 2A&B). In multivariate analysis right sided line placement was associated with a non-significant trend for increased CA-BSI (HR 1.40, 95%CI 0.99-1.96, p=0.055) and CR-BSI (HR 2.93, 95%CI 0.93-9.24), p=0.066).
Conclusion: In a randomized controlled trial the rate of CA-BSI was not impacted by whether CVAD insertion was on the dominant or non-dominant side. There is a suggestion, however, that a right sided line insertion may increase both CA-BSI and CR-BSI but a larger study would be required to properly assess this hypothesis.
No relevant conflicts of interest to declare.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal