Clinical Predictors of Infection of Central Venous Catheters Used for Total Parenteral Nutrition

Abstract

To identify predictors of infection in catheters used for total parenteral nutrition (TPN), clinical and microbiological data were prospectively collected on 169 catheter systems (88 patients). Based on semiquantitative catheter cultures, infection was associated with a positive insertion site skin culture taken close to the time of catheter removal (relative risk [RR] = 4.50), especially one yielding greater than or equal to 50 colonies of an organism other than coagulase-negative staphylococci. Infection was also associated with erythema at the insertion site greater than 4 mm in diameter (RR = 2.93). In a subset of 67 catheters for which blood cultures were obtained, infection was also associated with a positive peripheral venous blood culture (RR = 5.90) and a positive central venous blood culture obtained through the catheter (RR = 5.44). Based on a logistic regression model, periodic cultures of the insertion site should be useful in evaluating subsequent fever in stable patients with indwelling central venous catheters. Another source of fever is likely if inflammation is absent and there is either no colonization or there is colonization by less than 50 colonies of coagulase-negative staphylococci at the insertion site. Conversely, the catheter should be removed and cultured semiquantitatively if the site is colonized by an organism other than coagulase-negative staphylococci. We suggest that blood culture results add little to the risk estimate in these situations.

Diagnosis of infection in sepsis: An evidence-based review

“…Few studies have investigated skin insertion sites cultures in ICU patients (82). In patients with long-term nutrition catheters, negative site cultures had a negative predictive value for line infection of 98% (83), whereas positive culture, particularly of organisms other than coagulase-negative staphylococci, is predictive of CVC infection (79). In patients with nontunneled CVCs, the presence of Ͼ50 cfu of organisms by quantitative skin site cultures at the time of removal for suspected infection had a sensitivity of 75%, a positive predictive value of 100% and a negative predictive value of 92% in detecting CVC infection (84).…”

“…In patients with nontunneled CVCs, the presence of Ͼ50 cfu of organisms by quantitative skin site cultures at the time of removal for suspected infection had a sensitivity of 75%, a positive predictive value of 100% and a negative predictive value of 92% in detecting CVC infection (84). In another study, a site culture yielding Ͼ50 cfu coagulase-negative staphylococci was associated with a 4.4 times greater risk of catheter colonization (79). However, in one study, the positive predictive value of nonquantitative insertion site culture was only 29% (85).…”

“…For example, Safdar and Maki (78) found that among 1,263 central venous catheters, the index of local inflammation (including pain, erythema, swelling, and purulence) did not differ among noncolonized catheters, colonized catheters (n ϭ 333), and those associated with bacteremia (n ϭ 35), although the latter tended to have a higher index. In one study, erythema extending Ͼ4 mm around the skin entry site was found to be significantly associated with catheter infection (79). Therefore, while overt inflammation is suggestive of infection (68), local signs cannot be usually relied on to clinically identify catheter infection.…”

Novel Strategies to Prevent Catheter-Associated Infections in Oncology Patients

“…In a clinical study conducted by Collignon et al 85% of catheters colonized by >100 CFU (SQM) were associated with CRB 28 . Armstrong et al found a correlation for SQM >15 CFU on catheter tips and a colonization of the skin insertion site (>50 CFU) with CRI of catheters used for total parenteral nutrition 29 . Comparative quantitative blood cultures via CVC and peripheral blood with a 5-fold difference was significant for CRS 30 which has been widely accepted.…”

Central line infection: a review

“…However, most studies (Brendel, 1983;Messner, 1985) suggest that the distal 5 cm of the catheter be clipped off with sterile scissors and sent to the laboratory in a sterile container (not in culture medium) for culture. In most studies (Sitzmann et al, 1985;Maki & Ringer, 1987;Armstrong et al, 1990), it is generally agreed that there is a significant association between a positive skin culture and catheter infection. It was also generally agreed that when the patient showed signs of infection, blood cultures were done from the catheter and/or from a peripheral vein, the catheter was removed and the patient given antibiotics.…”