The Incidence of First Hickman Catheter-Related Infection and Predictors of Catheter Removal in Cancer Patients

Publisher Website
Google Scholar
Abstract
Objective: To describe the incidence and types of first Hickman catheter-related infection (HCRI) in cancer patients and to identify indicators for catheter removal. Design: Retrospective cohort study. Setting: A regional, tertiary, referral cancer center and its supportive care university teaching hospital. Patients and Methods: A retrospective review was conducted of 316 consecutive adult oncology patients who underwent Hickman catheter placement from 1986 to 1990 at a regional oncology center. HCRI was determined on the basis of clinical information incriminating the Hickman catheter as the source of infection. Patient characteristics and data about HCRIs (exit site cellulitis, tunnel infection with concomitant exit site cellulitis, bloodstream infection, and exit site cellulitis with bloodstream infection) were abstracted from patient medical records. Subsequently, univariate and multivariate analyses for the risk of HCRI and catheter removal were completed. Results: The incidence of first HCRI was 5.98 infections per 1,000 catheter days. Overall, 156 (49%) of 316 patients developed their first HCRI prior to catheter removal. The median time to HCRI was 90 days. Male gender (P = .0004) and hematologic malignancy (P = .0001) emerged as significant risk factors for HCRI in the univariate analysis. A Cox model verified that male gender (P = .02) and hematologic malignancy (P = .004) were associated with an enhanced risk of HCRI. There were 35 exit site infections (23%), three infections of the tunnel and the exit site (2%), 80 bloodstream infections (51%), and another 38 bloodstream infections with concomitant exit site infections (24%). The incidence of bloodstream infection was 3.05 per 1,000 catheter days. Gram-positive pathogens outnumbered gram-negatives and fungi, with Staphylococcus epidermidis being most common. Fifty (32%) of 156 HCRIs resulted in catheter removal. Predictors of Hickman catheter removal in the univariate analysis were bloodstream infection (P = .046) and pathogen type (P =.006). Multiple regression analysis suggested that having a gram-negative (P = .014) or fungal (P = .057) pathogen was the most important factor for catheter removal. Conclusions: These data suggest that first HCRIs occur more commonly in male patients with hematologic malignancies than in patients with solid tumors. The removal of Hickman catheters in oncology patients probably is predicated on the causative pathogen, but further investigations are necessary to delineate this issue.