3:45 PM Abstract No. 360 Predictors of catheter-related infections of tunneled lines in the inpatient setting

Abstract

Inpatients with central venous catheters (CVCs) that show signs of local or systemic infection may undergo CVC removal followed by catheter tip culture for confirmation of catheter-related infections (CRI) and catheter-related bloodstream infections (CRBSI), as defined by positive catheter tip culture with a congruent positive blood culture. The use of this test has dwindled over the past decade, owing to factors including reimbursement penalties, poor predictive value, and increased reliance on blood cultures (BCx). We aim to develop a clinical risk calculator to reduce unnecessary removal and catheter tip culture. This IRB-approved retrospective study examined 151 patients in the inpatient setting who underwent tunneled CVC removal and tip culture. The primary outcome was the presence of CRI as defined by a positive catheter tip culture. Subgroup analysis was performed to determine which clinical factors results influenced the presence of CRI. All analyses were performed using χ2 analysis, and binary logistic regression with backward elimination. In this cohort, the median age at placement was 52 years, and median catheter-days was 98 days. Of the 151 lines removed, 32 patients (21%) had CRBSI and 43(29%) of catheter tips were culture positive. Variables predictive of CRI were gram-positive BCx prior to line removal and access site purulence (OR = 2.2 and OR = 4.0, P < 0.05). Variables that were not predictive of CRI included: fungal BCx, access site and laterality, CVC type, white count, placement indication, catheter tip location, signs of systemic infection, tenderness, and erythema. While not in the final model, CVC cultures with gram-negative organisms were related to the absence of CRIs (χ2 value = 3.6, P < 0.05). In this population, it is important to discern high-risk patients. Our model demonstrates feasibility of a risk calculator in the setting of possible CRI. When choosing a course of action, clinicians may consider removing CVCs in the presence of purulence or gram-positive BCx, and conversely, treating through non-gram positive BCx infections. Limitations of the study include cohort size and antibiotics given before cultures.