Influence of the Insertion Site on Central Venous Catheter Related Complications in Patients Undergoing Allogeneic Hematopoietic Cell Transplantation, High Dose Therapy or Induction Therapy for Acute Leukemia

Central venous catheters (CVCs) are extensively used in patients undergoing hematopoietic cell transplantation (HCT) or induction chemotherapy for acute leukemia. In these patients CVCs are placed routinely either via internal jugular (IJV) or subclavian veins (SCV). Several studies have compared rates of complications between both insertion sites, particularly in the intensive care unit setting. However, to our knowledge there are no data in the literature comparing the insertion sites IJV versus SCV in patients undergoing allogeneic HCT or high dose therapy for hematologic malignancies. Thus, it was the purpose of this study to analyze systematically complications of CVCs at the different insertion sites in these patients.

Patients and methods:All consecutive patients who were treated in the unit for adult HCT at our institution from 01/2011 to 06/2013 were included into this retrospective analysis. Inclusion criterion was insertion of a CVC due to allogeneic HCT, autologous HCT or induction therapy for AML or ALL. 3-lumen standard CVCs (Arrow, Reading, PA) were placed after informed consent either via the IJV or the SCV. CVCs were inserted by experienced physicians after disinfection by 70% propanolol under full barrier precautions. The insertion site was chosen at the responsible physician´s own discretion. CVCs were covered by chlorhexidine gluconate impregnated dressings, either as transparent polyurethane (Tegaderm CHG, 3M, Saint Paul, MN) or as sponge dressing (Biopatch, Ethicon, Cincinnati, OH). No patient received a systemic antibiotic prophylaxis. Reasons for premature removal of CVCs were suspected or proven CVC-related blood stream infection (BSI), progression of local infections at the insertion site, or any other CVC related severe adverse event. End points of this study were local infections, fever, BSI, duration of catheterization per CVC, congestion of the CVC, thrombosis, pneumothorax and suspected or proven catheter related deaths.

Results: During the study period 255 CVCs were placed in 170 consecutive patients [median age: 59 years (range 19-79), F: n=63; M: n=107] due to allogeneic HCT (65 patients, 101 insertions), autologous HCT (30 patients, 33 insertions) or induction therapy for AML or ALL (75 patients, 121 insertions). Underlying diseases were AML (n=102), ALL (n=9), MDS (n=15), MPN (n=4), lymphoma (n=21) or plasma cell neoplasia (n=19). Out of the 255 CVSs 155 (61%) were sited via the IJV and 100 (39%) via the SCV. Among the 101 CVCs in the allogeneic HCT group 60 (60%) were placed in the IJV and 41 (41%) in the SCV. The median duration of catheterization/per CVC for the entire group was 16 days (range 3-49, IJV: 17, SCV: 15 days, ns). The corresponding data for the allogeneic HCT group were 18 (IJV) and 17 day (SCV). The events fever or local infection occurred in 75% (45/60) of IJV- and in 95% (39/41) of SCV-CVCs (p=0.0084) in the allogeneic HCT group. Moreover, in SCV-CVCs these events arose significantly earlier (median time to event 7 vs. 10 days in the IJV (p=0.012). Focusing only on the event local infection the results were also superior for the IJV-CVCs: 28/60 (47%) vs. 29/41(71%) for SCV-CVCs (p=0.026). For the entire cohort of CVCs data were similar: The events fever or local infection occurred in 127/155 (82%) of IJV- and 95/100 (95%) of SCV-CVCs (p=0.025). Median time to these events were 8 (IJV) vs. 10 days (SCV, p=0.0009). For local infections alone the differences were even more pronounced: 81/155 (IJV: 52%) vs. 81/100 (SCV: 81%, p<0.0001). There were no significant differences in the frequency of BSI between the insertion sites [entire group: IJV 46/155 (30%), SCV 20/100 (20%); allogeneic HCT group: IJV 14/60 (23%), SCV 12/41 (29%)]. Thrombosis or pneumothorax occurred only once, both with SCV-CVCs. Congestion of CVCs occurred in 20% (31/155) IJV- and in 24% (24/100) of SCV-CVCs (ns), respectively. Data for allogeneic HCT were 25% (15/60) in IJV- and 27% (11/41) in SCV-CVCs (ns). Out of the entire cohort 10 patients (7 x allogeneic HCT, 3 x induction therapy) died with an inserted CVC. 5 of them died due to BSI and therefore potentially CVC-related (3 x IJV, 2 x SCV).

Conclusions: In contrast to earlier studies on patients without hematologic malignancies these data demonstrate clearly that CVCs placed in the SCV are not superior. Moreover, local infections and fever occurred significantly earlier and more frequent in SCV-CVCs in patients undergoing allogeneic HCT.