Portal de Periódicos da CAPES

Conditions associated with infections of indwelling central venous catheters in cancer patients: a summary

2003; Wiley; Volume: 121; Issue: 2 Linguagem: Inglês

10.1046/j.1365-2141.2003.04209.x

1365-2141

Elio Castagnola, Angelo Claudio Molinari, Giuzeppe Fratino, Claudio Viscoli,

Acute Kidney Injury Research

The aim of the present annotation was to summarize the clinical situations that are related to the risk of the development of infections of indwelling central venous catheters (CVCs). The specific catheter type (partially implanted), the underlying disease and its management (acute leukaemia), and the patient's age (paediatrics) are related to more frequent CVC manipulations, and represent the most important risk factors for infection. This risk is clearly reduced by the use of totally implanted CVCs and the proficient training of the individuals performing the CVC maintenance procedures. The CVC insertion technique (percutaneous versus surgical insertion or radiological suite versus operating theatre) does not seem to modify the risk of infection. Finally, there is disagreement as to whether particular therapeutic procedures, such as bone marrow transplant, and the presence of neutropenia are risk factors associated with a higher incidence of CVC-related infections. Indwelling central venous catheters (CVC) are an important tool in the management of patients receiving antineoplastic chemotherapy or a bone marrow transplant, but are also associated with the risk of infectious complications (Viscoli & Castagnola, 1995). Contamination of the catheter hub and lumen appear to be the predominant cause of CVC-related bacteraemia development (Weightman et al, 1988; Pascual, 2002). Recently published reviews have clarified the case definition, aetiology and therapeutic options of CVC-related infections (Mermel et al, 2001; Bouza et al, 2002; Caratallà, 2002; Paiva & Pereira, 2002; Rodriguez-Baño, 2002). Therefore, these topics will not be covered in this article. As the clinical situations related to the risk of developing CVC-related infections are usually described in a fragmented fashion, the aim of the present article was to summarize all the conditions that are commonly associated with the risk of developing these complications. The clinical studies and observations regarding many of these conditions are reported in studies with limited strength of evidence. Thus, no ‘ranking’ of evidence will be given, and we will restrict our work to a report of the clinical conditions that have been observed and the inferences that can be made from them. To date, there are no studies demonstrating an association between the indwelling CVC's material and the risk of infection. However, in general, a higher bacterial adherence to siliconized latex and polyvinyl chloride catheters has been shown when compared with catheters made of other materials (Pascual, 2002). Indwelling CVCs used for the management of underlying disease in cancer patients can be mainly divided into two groups: (1) totally implantable catheters [so called ‘ports’ or totally implantable venous access devices (TIVADs)], which are either valved or non-valved, and (2) partially implantable, or tunnelled, catheters (so called ‘Hickman/Broviac’ CVCs), which can be further divided into (a) non-valved (‘true’ Hickman–Broviac type), (b) with a valved tip (so-called ‘Groshong’ type) and (c) pressure-activated safety-valved catheters (PASVs). All devices (either totally or partially implanted) may have one or more lumens or reservoirs. The main difference between the various types of CVCs is the number of maintenance procedures required to prevent catheter obstruction when not in use: 2–3 times a week for Hickman–Broviac, once a week for valved catheters and once every 4 weeks for TIVADs. Moreover, the cutaneous emergence of partially implanted CVCs requires maintenance procedures that are obviously absent in totally implanted ports (Mermel, 2000). Totally or partially implanted CVCs. Two recent reviews, which included studies performed on both children and adults, summarized the differences in infection rates between totally and partially implanted CVCs. In the first review (Crnich & Maki, 2002), the authors analysed the results of a number of prospective studies on infectious complications occurring in long-term CVCs, and showed that the pooled mean number of CVC-related bacteraemias per 100 catheters was 5·1 [95%confidence interval (CI) 4·0–6·3] in 13 prospective studies involving TIVADs and 20·9 (95%CI 18·2–21.0) in 18 prospective studies of tunnelled CVCs. Similar differences were observed in the pooled mean incidence of the same complication per 1000 catheter d, which were found to be 0·2 (95%CI 0·1–0·3) in TIVADs and 1·2 (1·0–1·3) in Hickman–Broviac CVCs. In the second review (Henderson, 2000), which reported data from 20 studies, the mean incidence of CVC-related bacteraemias was 11·2% in TIVADs and 36·7% in Hickman/Broviac CVCs (P = 0·008). Moreover, the mean incidence rate of CVC-related bacteraemias in totally implanted CVCs (0·059 per 1000 catheter d) was found to be significantly lower (P = 0·001) than that of partially implanted CVCs (3·9 per 1000 catheter d). In the same review (Henderson, 2000), the mean incidence of exit-site/tunnel-pocket infections was 7% in TIVADs and 15·8% in tunnelled CVCs, a difference found not to be statistically significant (P = 0·33). However, the incidence of exit-site/tunnel-pocket infections is not often reported and incidence rates vary between studies. A previous study, regarding 690 Hickman–Broviac catheters in adults with neoplastic disease (Newman et al, 1993), reported 23% exit-site infections and 7% tunnel infections, with an incidence rate of 1·53 per 1000 catheter d. However, a recent prospective analysis of tunnelled CVCs in children (Fratino et al, 2002), showed an overall 4% exit-site/tunnel infection rate, with an incidence of 0·22 episodes per 1000 catheter d. On the other hand, pocket infections in patients with TIVADs are very rare events. Freytes et al (1990) observed a 1·5% rate of pocket infections, in a study involving 134 TIVADs in adult patients, and Biffi et al (1997) observed a 0·56% rate of pocket infections with an incidence of 0·03 per 1000 catheter d, in a study involving 178 TIVADs in adults. Valved catheters were introduced in clinical practice in an attempt to reduce the number of manipulations required when not in use. Although only few data regarding infectious complications in valved CVCs are available, it would seem that the introduction of these devices has not significantly changed the occurrence of infectious complications. In fact, in a study involving 104 CVCs in adults, Gleeson and co-workers observed 32% of CVC-related bacteraemias in patients fitted with Groshong-type CVCs, compared with 16·2% in patients with Hickman-type CVCs (Gleeson et al, 1993). On the other hand, Keung and co-workers did not observe any significant differences in the incidence of complications in a retrospective study of 111 CVCs in similar patients, including Groshong, Hickman and TIVAD CVCs (Keung et al, 1994). Lastly, a prospective, non-randomized comparison between Groshong and Hickman catheters performed on a paediatric population did not demonstrate a statistically significant difference in the incidence of infectious CVC-related complications between the two CVC types (Biagi et al, 1997). There are even fewer data regarding infection rates of PASV catheters. In a prospective, non-randomized study comparing the incidence of complications between PASV- and Broviac-type CVCs in children either with cancer or receiving bone marrow transplant, the incidence of infectious complications was similar in the two groups: 12% of both PASV and Broviac CVCs types presented with infectious complications, with incidence rates of 0·9 and 0·5 per 1000 catheter d respectively (Fratino et al, 2002). The presence or absence of a Groshong-type valve in TIVADs also did not alter the incidence of infectious complications. In fact, Biffi et al (2001) published a randomized study performed on adults that compared valved and non-valved TIVADs, which showed a very low overall incidence of infectious complications (less than 1%), even though valved TIVADs did not prove to be superior to ‘classical’ catheters in terms of early and late complications. It has been suggested that the number of CVC lumens (single or double) is an important factor in the development of CVC-related infections. In a retrospective study regarding 91 double-lumen and 51 single-lumen Hickman catheters in adults (Early et al, 1990), the infection rate was observed to be 2·02 per 1000 catheter d in double-lumen CVCs and 0·83 per 1000 catheter d in single-lumen CVCs. A trend towards a higher incidence of complications in double-lumen CVCs was observed in another retrospective study involving 111 single- or double-lumen catheters of different types (Keung et al, 1994). Manipulation of the catheter is an important risk factor for the development of infectious complications. The number of manipulations, the ability of the people performing the procedures in either the hospital or at home, the materials employed for maintenance procedures, the patients' socio-economic conditions and the correct functioning of the catheter are all factors that could represent a risk for CVC-related infections. The number of CVC manipulations is determined by both the type of catheter and by the patient's underlying disease (as well as the management of chemotherapy-related complications), and is closely related to the risk of bacteraemia. In a study comparing the number of CVC-related infections in adults with acute leukaemia versus those with solid tumours, it was demonstrated that patients with leukaemia had a comparatively greater number of CVC manipulations for the management of their underlying disease and, therefore, a higher incidence of CVC-related bacteraemias (Duthoit et al, 1993). Multivariate analysis to identify the risk factors associated with CVC-related infections in adults with acute leukaemia has demonstrated that CVC use and neutropenia were independent risk factors for bacteraemia development (Pagano et al, 1997). Moreover, in a cohort of adult patients with cancer admitted to an intensive care unit, it was demonstrated that CVC-related infections developed more often in the patients with more severe illness and who stayed in the intensive care unit longer, which are both conditions that are associated with more frequent CVC manipulations (Velasco et al, 1997). Therefore, any attempt to reduce the ‘routine’ number of CVC manipulations at any site could reduce the risk of CVC-related infections. It has been demonstrated that infusion sets can safely be replaced every 72 h, except when following the infusion of blood products and lipid emulsion administration (Pearson, 1996). Two studies performed on adults undergoing bone marrow transplants showed that fewer CVC dressing changes, i.e. a reduced number of manipulations, decreased the incidence of exit-site/tunnel infections in partially implanted CVCs (Shivnan et al, 1991; Rasero et al, 2000). It is commonly stated that a malfunctioning catheter has a higher probability of becoming infected but this has rarely been analysed. An autopsy study performed on adults with venous thrombosis after CVC placement demonstrated that 22% of these patients had a previous CVC-related infection (Raad et al, 1994). However, neither a retrospective study performed on 567 CVCs inserted in children with cancer (Fratino et al, 2001; Molinari et al, 2001) nor a prospective evaluation of 92 CVCs in a similar population (Fratino et al, 2002), showed a strict correlation between malfunctioning of the catheter and the development of subsequent infectious complications. Good training in catheter manipulation procedures is probably one of the best methods for guarding against CVC-related bacteraemias both in the hospital and at home (Rizzari et al, 1992; Daghistani et al, 1996; Castagnola et al, 1998; Abi-Said et al, 1999; Shah et al, 2002). The presence of a specialized team of nurses committed to the care of CVCs (Mermel, 2000; Eggimann & Pittet, 2002) or the knowledge of proper maintenance procedure techniques shared by all nurses (Abi-Said et al, 1999; Eggimann & Pittet, 2002) are both effective approaches for the reduction of CVC-related infection incidence in hospitalized patients. Non-medically trained individuals frequently perform a great number of catheter maintenance procedures at home. These individuals play an important role in determining the risk of infectious complications in indwelling CVC carriers (Rizzari et al, 1992; White & Ragland, 1994; Castagnola et al, 1997; Castagnola et al, 1998). Therefore, it is necessary to maintain a close microbiological surveillance and to also promote educational programmes in these settings. Contamination of infusates used for CVC maintenance procedures, especially by Gram-negative bacteria, represents an important cause of ‘epidemics’ of CVC-related bacteraemias (Bozzetti et al, 1990; Castagnola et al, 1994, 1998; Namnyak et al, 1999; Playford et al, 1999). Poor hygiene (Castagnola et al, 1997) and low socio-economic conditions, such as those seen in developing countries (Ertem et al, 1999; Burney et al, 2001), have been associated with an increased risk of indwelling CVC-related infections. Moreover, cultural, ethnic and language differences have been suggested as risk factors for developing CVC-related infections in children with cancer because of a poor understanding of CVC care training by parents (Kellerman et al, 1996). As already stated, patients with acute leukaemia require a greater number of CVC manipulations for the management of chemotherapy-induced side-effects and complications, thus causing a higher incidence of CVC-related infections when compared with patients with solid tumours (Duthoit & Devleeshouwer, 1993; Pagano et al, 1997). It is commonplace to state that the incidence of CVC-related infections is higher in children, who frequently require a greater number of CVC manipulations (Decker & Edwards, 1988), than in adults. However, to our knowledge there are no studies analysing the incidence of CVC-related infectious complications in patients of differing ages but with similar underlying disease. Bone marrow transplant. The role that bone marrow transplants play in CVC-related infections is controversial. The transplant procedure during the hospital stay has been considered to be a risk factor for CVC infection per se in studies involving both children and adults (Uderzo et al, 1992; Keung et al, 1995). However, this observation was not confirmed in a more recent paediatric report where CVC-related infections were more frequently observed as late complications (more than 100 d from the procedure) of allogeneic bone marrow transplantation (Romano et al, 1999), probably as a result of the greater number of CVC manipulations required for procedures directed to control graft-versus-host disease. Neutropenia. Antineoplastic chemotherapy-induced granulocytopenia represents one of the most important risk factors for infections either in cancer patients or following bone marrow transplant, but its association with indwelling CVC-related infection development is controversial, as its involvement has been suggested by some reports (Biagi et al, 1997; Pagano et al, 1997), but has not been confirmed by other studies performed both on children and adults (Viscoli et al, 1988; Gorelick et al, 1991; Castagnola et al, 1994, 1995; Romano et al, 1999). However, it has been suggested that neutropenia at the time of catheter insertion represents an important risk factor for CVC-related infections, with an increase in the relative risk of infectious complications (up to 20-fold) observed, leading to CVC removal (Groeger et al, 1993; Shaul et al, 1998; Nouwen et al, 1999). Skin preparation, maximum use of sterile barrier precautions during insertion, and the rigorous cleaning and disinfection of the insertion site are pivotal for the reduction of CVC-related infection risks associated with catheter insertion (Eggimann & Pittet, 2002). Technique of insertion: percutaneous versus surgical insertion. In a study comparing percutaneous versus surgical insertion for Hickman (n = 120), Broviac (n = 146) or port (n = 93) CVCs in adult cancer patients, percutaneously inserted catheters developed infections comparatively less frequently, even though these complications were not strictly linked to the CVC insertion (Mirro et al, 1990). In another study comparing the two insertion techniques in a similar population, 19% of bacteraemias, 19% of tunnel infections and 7% of exit-site infections were observed after percutaneous insertion vs 41% of bacteraemias, 18% of tunnel infections and 26% of exit-site infections after surgical insertion (Ahmed & Mohyuddin, 1998) In addition, in a study performed on 129 TIVADs positioned in adults, 3% of catheters inserted percutaneously presented a CVC-related infection vs 0% of catheters that were inserted surgically (Puig-la Calle et al, 1996). No major infectious complications were associated with vein cutdown approach in 100 CVCs inserted in adults with cancer (Povoski, 2000). Lastly, in a retrospective study on 305 indwelling central venous catheters, Minassian and co-workers observed that the ‘cut-down’ approach in the operating theatre was associated with a lower complication rate than the percutaneous approach (Minassian et al, 2000). Location of catheter insertion. No significant variation in the risk of infection has been observed in relation to the hospital location of catheter insertion, i.e. operating theatre versus radiology suite (Eastridge & Lefor, 1995; Nouwen et al, 1999). Site of catheter insertion. Two studies, which included patients of different ages, support the observation that catheters inserted in the femoral vein present a higher incidence of infectious complications. In one study, catheters placed in the vena cava area presented 16·7% of CVC-related bacteraemias, compared with 1·8% in CVCs positioned in the superior vena cava area (Harden et al, 1995). In another study, 15% of CVC-related infectious episodes were observed in catheters inserted in the inferior vena cava area (Haire et al, 1995). In contrast, femoral access resulted in safe catheter insertion without any major infectious complications in a small study of children in whom access to the veins of the upper part of the torso was difficult or contraindicated (Sovinz et al, 2001). In all patients, the long tunnelling of CVCs with peri-umbilical emergence is suggested as the best procedure for reducing CVC-related infections. Unfortunately, few studies have specifically analysed this topic, and definitions used for these purposes were found to be different and non-comparable. Boulay and co-workers reported a 7·7% rate of surgical wounds infection following the placement of 78 CVCs in women with gynaecological malignancies, but these infections were mainly associated with chemotherapy administration after catheter insertion (Boulay et al, 1998). In a prospective study of tunnelled CVCs in children with cancer, 2% of patients presented with an infectious episode (one tunnel infection and one exit-site infection) within the first 10 d after catheter positioning (Fratino et al, 2002). In another study involving 473 Hickman–Broviac catheters in adults, the authors reported a 12% rate of wound infections within 45 d of CVC insertion (Shaul et al, 1998). Finally, 1% of insertious wound infections (TIVADs 1·5% Hickman-Broviac 0%) were observed within 30 days from insertion in a prospective study on 100 CVCs in adults (Povoski, 2000). In these settings, the highest risk of infection was observed in neutropenic patients, but the risk was reduced if patients received antibacterial prophylaxis at the time of CVC insertion. It must be stressed that, to date, no prospective, randomized studies have demonstrated that antibiotic administration during insertion plays a role in the prevention of CVC-related infections, as has recently been reviewed (Mermel, 2000). Infections in indwelling central venous catheters are determined by many different factors. The number of CVC manipulations represents the most important risk factor for the development of CVC-related infections. Catheter type (partially implanted), underlying disease and its management (acute leukaemia), and the patient's age (paediatrics) are factors requiring more frequent CVC manipulations, and are, therefore, associated with an increased risk of developing infectious complications. This risk is clearly reduced by the use of totally implanted CVCs (when feasible) and proficient training of the individuals performing CVC manipulations, both in hospital and in the patients' home. The CVC insertion technique (percutaneous versus surgical insertion or radiological suite versus operating theatre) does not seem to modify the risk of CVC-related infections, even though the particular aspect of catheter positioning, wound infections, as well as the role of catheter patency or malfunctioning, has rarely been addressed. Lastly, it is interesting to note that controversy exists as to whether or not particular therapeutic procedures, such as bone marrow transplant, and the presence of neutropenia are factors associated with a higher incidence of CVC-related infections. The authors thank Mrs Valerie Perricone for reviewing the text.