Original Paper Blood Purif 2009;28:21–28 DOI: 10.1159/000210034 Received: February 1, 2008 Accepted: October 15, 2008 Published online: March 27, 2009 Analysis of Risk Factors for CatheterRelated Bacteremia in 2000 Permanent Dual Catheters for Hemodialysis Xavier Lemaire a Marion Morena b Hélène Leray-Moragués a Delphine Henriet-Viprey a Leila Chenine a Christine Defez-Fougeron c Bernard Canaud a, b a Service de Néphrologie, Hémodialyse et Soins Intensifs, CHU Lapeyronie, and b Institut de Recherche et de Formation en Dialyse, Montpellier, and c Département Informatique Médicale, CHU Caremeau, Nîmes, France Key Words Bacteremia ⴢ Catheter ⴢ Hemodialysis ⴢ Infection Abstract Background: Infection constitutes a leading cause of morbidity and mortality in hemodialysis (HD) patients. The type of vascular access is an important determinant of the risk of infection. Therefore, identification of risk factors leading to catheter-related bacteremia (CRB) is strongly required. The aim of this prospective large cohort study of HD patients using only catheters as vascular access was to isolate risk factors for CRB. Methods: 2,230 permanent silicone dual catheters implanted in 1,749 patients between November 1982 and November 2005 were studied. The following data were collected at the time of catheter implantation: presence of hypertension, diabetes mellitus, obesity, atherosclerosis, immunodepression, Wright-Khan index, site and side of catheter insertion, and history of bacteremia. Results: The site of catheter insertion was internal jugular (n = 2,133), subclavian (n = 79) and femoral (n = 17). Duration of catheter use was as follows: 30–90 days (n = 1,607) and 190 days (n = 1,054); 226 episodes of bacteremia occurred in 197 catheters. Microorganisms responsible were mainly Staphylococcus aureus, co- © 2009 S. Karger AG, Basel 0253–5068/09/0281–0021$26.00/0 Fax +41 61 306 12 34 E-Mail karger@karger.ch www.karger.com Accessible online at: www.karger.com/bpu agulase-negative staphylococci, Enterobacter spp. and Pseudomonas aeruginosa. The overall incidence of bacteremic episodes was 0.514/1,000 catheter days. Hypertension, atherosclerosis, diabetes mellitus, site of catheter implantation, duration of catheter use, Wright-Khan comorbidity index and previous history of CRB were significant risk factors associated with bacteremia in univariate analysis. Multivariate analysis revealed that a previous history of a bacteremic episode (odds ratio, OR = 2.70, 95% confidence interval, CI = 1.56–4.68), diabetes mellitus (OR = 2.37, 95% CI = 1.65–3.39), duration of catheter use 190 days (OR = 1.85, 95% CI = 1.35– 2.55) and hypertension (OR = 1.49, 95% CI = 1.08–2.04) were still significant factors associated with bacteremia. Conclusion: Reducing CRB is still a challenge for nephrologists to reduce patient morbidity and mortality. Our study could demonstrate that diabetes, previous history of CRB, site of catheter implantation and duration of catheter use were the most important risk factors for bacteremia. Therefore, to prevent CRB, particular attention should be paid to patients with diabetes and a previous history of bacteremia following strict hygienic and aseptic rules for catheter handling associated with the regular use of antiseptic lock solutions. Copyright © 2009 S. Karger AG, Basel Prof. Bernard Canaud Service de Néphrologie, Hémodialyse et Soins Intensifs, CHU Lapeyronie 371, Avenue du Doyen Gaston Giraud FR–34295 Montpellier Cedex 5 (France) Tel. +33 467 338 955, Fax +33 467 603 783, E-Mail b-canaud@chu-montpellier.fr Introduction Infection constitutes a leading cause of morbidity and the second leading cause of mortality among hemodialysis (HD) patients [1, 2]. The type of vascular access is an important determinant of the risk of infection as well as mortality [3–5]. Indeed, use of central venous catheters and grafts has been associated with a markedly increased risk of death from sepsis [1] compared to arteriovenous fistulas (corresponding to a 7- and 3-fold increased risk for catheters and grafts, respectively). Since catheters represent the vascular access in 1.3 (Japan) to 28.2% (USA) of prevalent HD patients (data issued from DOPPS) [6], they still constitute a main source of infection in this population. Therefore, identification of risk factors for catheter-related bacteremia (CRB) is strongly required. Different studies [7–9] tried to isolate risk factors in order to prevent CRB. However, none of these studies were really standardized, and variable incidence rates have been reported (0.5–9/1,000 catheter days), with heterogeneous results concerning risk factors and the type of vascular access, and quite limited observation periods. Therefore, the aim of this 23-year prospective cohort study in patients with catheters (n = 2,230) as permanent vascular access was to identify risk factors for bacteremia possibly affecting HD care and practices. Case Definition Only bacteremic episodes usually detected after clinical symptoms including fever and chills, and defined as at least one positive blood culture were considered in the study, except for coagulase-negative staphylococci, Bacillus spp. and Corynebacterium spp., for which a single positive blood culture was considered as a contamination. Data Collection Between November 1982 and November 2005, 2,622 catheters were inserted in 2,095 patients. Retrospectively, the following data were collected at the time of catheter insertion: gender; age; weight, height and body mass index; cause of renal failure; dialysis modality; presence of hypertension, diabetes mellitus, obesity (body mass index 125), atherosclerosis (i.e. a history of clinical events), immunodepression (corticoids, vasculitis, connective tissue disease, multiple myeloma, liver cirrhosis, HIV, neoplasm, current immunosuppressive chemotherapy or nephrotic syndrome), comorbidity score using the Wright/Khan index (which is a three-point score from low, medium, to high comorbidity and takes patient age, presence of diabetes, malignancy and organspecific diseases into account) [11, 12], site, side and date of catheter insertion; history of bacteremia or another infection related or not to catheter use, and type of anesthesia. When a bacteremic episode was diagnosed, the number of positive blood cultures as well as the identification of microorganisms was recorded. All other non-infectious complications were also collected. At the end of catheter lifetime, the date of catheter removal, duration of catheter use, cause of removal [e.g. availability of another vascular access, catheter-related bacteremic infection, other non-infectious problems, death due to (i) CRB or (ii) other cause, transplantation, renal function recovery, or cessation of therapeutic treatment], number of bacteremic episodes and noninfectious problems during catheter use were also recorded. Patients and Methods Study Design Our study is a retrospective multicenter cohort of HD patients (from 4 different dialysis units) using only catheters as vascular access. All catheters, which have been inserted by five permanent senior nephrologists in the University Hospital Center of Montpellier, were all tunnelled silicone non-cuffed dual catheters (DCath) as presented in figure 1 (DualCath쏐; Hemotech Ramonville, France, and MedComp, Harleyville, Pa., USA) and implanted percutaneously according to a standardized method previously described [10]. Indication of DCath was as follows: insertion to permit creation, maturation or revision of a graft or a fistula; an exhausted vascular access site, or comfort for elderly and desperate patients. Twenty percent of all vascular accesses were permanent catheters (5% being PTFE grafts and 75% native arteriovenous fistulas). All DCath were inserted from November 1982 to November 2005. Only catheters with complete lifetime follow-up were selected; catheters with incomplete data or lost to follow-up were excluded. Finally, patients undergoing home and self-care HD were excluded from the study. At the catheter punction site, blood was sampled, and monthly routine blood cultures were performed in 25% of patients. 22 Blood Purif 2009;28:21–28 Catheter Handling Catheter Connection. The nurse must wear a mask, protection glasses and a cap; the patient must also wear a mask. After removing the old dressing, the skin and catheter site are disinfected using Betadine쏐 scrub followed by rinsing of the catheter site with 0.9% NaCl and finally drying. Then, the nurse wearing a coat and sterile gloves creates a sterile field under the catheters, removes the access cap with a Betadine swab, attaches an empty syringe directly to the catheter lumen, withdraws 5 ml of blood and checks for the presence of a blood clot. After rinsing of the catheters with 20 ml of 0.9%NaCl and injection of anticoagulant, the catheters are connected to the arterial and venous lines and the dialysis session starts. A protective dressing is made at the catheter site. Catheter Disconnection. Catheter disconnection was performed following the same hygienic and aseptic rules as for catheter connection. Regarding anticoagulant lock solution, protocols [mainly heparin, sodium citrate and TaurolockTM (containing 1.35% taurolidine, an antimicrobial agent, and 4% citrate solutions)] are adapted to the patient’s characteristics. In case of suspected infection, antibiotic therapy is prescribed given the fact that biofilm is the major source of CRB. The first-line treatment uses vancomycin, an optimal empiric systemic antibiotic regimen for suspected CRB, and Taurolock as an antimicrobial catheter lock solution. Of note, in contrast to certain other antibiotics, Lemaire et al. Taurolock does not result in the development of resistance. In the second-line treatment, antibiotic therapy is adjusted to the germ isolated. Finally, a tight dressing is applied at the catheter site. Dialysis Adequacy Dialysis adequacy was performed monthly by checking a dialysis dose urea Kt/V 11.2 and by measuring standard parameters (e.g. Ca/P metabolism or hemoglobin) in agreement with clinical practice guidelines. Statistical Analyses Data were expressed as frequencies (percentages) for qualitative variables, means + SD for normally distributed variables and as medians (min–max) for non-normally distributed variables. Qualitative parameters were compared by ␹2 test or Fisher’s exact test if expected frequencies are !5. Quantitative parameters were compared using Student’s t test or the Mann-Whitney-Wilcoxon test according to the distributions. Univariate logistic regression was conducted for all predictor variables. Variables that were significantly associated with the presence of bacteremia at a p value ^0.2 were included in multivariate analysis, which was used to assess the relative importance of the different risk factors associated with the presence of bacteremia episodes. The results of the statistical tests are presented as odds ratios (OR) with their 95% confidence intervals (CI) and the degree of significance (p value). The appropriateness of the model was checked using the Hosmer-Lemeshow goodness-of-fit test. Finally, a survey study using the Kaplan-Meier method was performed to determine the median time of DCath survival. The survival rate of patients, comparing those who experienced bacteremia with those who did not, was also subjected to KaplanMeier survival analysis and log-rank testing. A p value !0.05 was considered to be statistically significant. All analyses were carried out with SAS software, version 8.1 (SAS Institute, Cary, N.C., USA). Fig. 1. Typical aspect of DCath inserted in HD patients after few weeks of catheter use. Table 1. Causes of catheter ablation Cause of catheter ablation Catheter Proportion, % Other functional vascular access CRB Death related to bacteremia episode Non-infectious event of the catheter Death due to other causes Transplantation Cessation of treatment Renal function recovery 1,503 62 7 34 369 34 16 205 67.40 2.78 0.31 1.52 16.55 1.52 0.72 9.20 Results Patients In the 2,095 patients initially included in the study, the causes of renal failure were as follows: glomerulonephritis (n = 507), nephroangiosclerosis and hypertensive nephropathy (n = 402), chronic interstitial nephritis (stone disease or infection; n = 127), hereditary renal disease (autosomal polycystic kidney disease or Alport syndrome; n = 121), diabetes mellitus (n = 184), myeloma, hemopathy and amyloidosis (n = 63), chronic interstitial and tubular nephritis (ischemic and toxic nephropathy; n = 192), unknown causes (n = 674) or other causes (n = 352). From these 2,095 patients, 346 non-censured patients (corresponding to a total of 391 catheters) had not been taken into consideration (due to recent insertion of a catheter). As only catheters with complete data were analyzed, the statistical analysis was performed on the remaining 1,749 patients corresponding to a total of 2,230 Risk Factors for Catheter-Related Bacteremia catheters, each catheter being taken into consideration as one studied case. Characteristics of the patients are listed in table 1. Median age of the patients was 62.8 years (range, 16–93). Description of Bacteremic Episodes A total of 226 episodes of bacteremia occurred in 197 catheters (with the occurrence of 1 episode in 171 cases, 2 in 11 cases, 3 in 9 cases and 6 in 1 case). As shown in figure 2, the microorganisms responsible for bacteremia were mainly Staphylococcus aureus (n = 106, 48%), coagulasenegative staphylococci (n = 35, 15%), Enterobacter spp. (n = 21, 9%) and Pseudomonas aeruginosa (n = 21, 9%). Of the 226 episodes of bacteremia reported, 42 were complicated by local or secondary septic focus complications: spondylodiscitis (n = 3), endocarditis (n = 1), meningitis (n = 1), arthritis (n = 1), osteitis (n = 1) and abscess or tunnelitis (n = 35). In 7 cases (0.31%), bacteremia was fatal. Blood Purif 2009;28:21–28 23 9% (21) 2% (5) 15% (35) E. coli Enterobacter spp. Klebsiella pneumoniae Pseudomonas spp. Serratia spp. Staphylococcus aureus Coagulase-negative staphylococci Other 9% (21) Incidence rate (/1,000 catheter-days) Fig. 2. Distribution (%) of bacteria responsible for episodes of bacteremia. A total of 226 episodes of bacteremia occurred in 197 catheters (with the occurrence of 1 episode in 171 cases, 2 in 11 cases, 3 in 9 cases and 6 in 1 case). Microorganisms responsible were mainly S. aureus (n = 106, 48%), coagulase-negative staphylococci (n = 35, 15%), Enterobacter spp. (n = 21, 9%) and P. aeruginosa (n = 21, 9%). 3% (7) 48% (106) 0.6 Wright-Khan index (2 vs. 0) 0.5 Wright-Khan index (1 vs. 0) 0.4 Duration of catheter use (>90 days vs. < 90 days) 0.3 Diabetes mellitius (yes vs. no) 0.2 Obesity (yes vs. no) 0.1 Atherosclerosis (yes vs. no) 0 Total 1982– 1995 Immunodepression (yes vs. no) 1995– 2005 Hypertension (yes vs. no) 0.5 1 2 3 4 Odds ratio Fig. 3. Mean annual incidence of bactere- mia compared to two periods (1982–1995 and 1995–2005) showing that there were no significant differences during the study. Fig. 4. Univariate risk analysis of factors associated with an increased risk of bacteremia with DCath. Incidence of Bacteremic Episodes The number of days of catheter use was 439,525 (with a mean duration of 6.6 months). As 226 episodes of bacteremia occurred, the overall incidence of bacteremic episodes was 0.514 per 1,000 catheter days, this incidence rate being stable over time (era 1982–1995: 0.49 per 1,000 catheter days; era 1995–2005: 0.53 per 1,000 catheter days; fig. 3). Causes of Catheter Removal Of the 2,230 catheters inserted, causes of removal were as follows: other functional vascular access (n = 24 Blood Purif 2009;28:21–28 1,503, 67.40%), catheter-related bacteremic episode (n = 62, 2.78%), death related to bacteremia (n = 7, 0.31%), non-infectious event of the catheter (n = 34, 1.52%), death due to other cause (n = 369, 16.55%), transplantation (n = 34, 1.52%), cessation of treatment (n = 16, 0.72%) and late incomplete renal function recovery (n = 205, 9.20%). Risk Factors for Bacteremia The univariate analysis clearly showed that hypertension (OR = 1.86; 95% CI = 1.38–2.52; p ! 0.0001), atherosclerosis (OR = 1.77; 95% CI = 1.30–2.42; p ! 0.0001), Lemaire et al. Color version available online 5% (11) 9% (20) Table 2. Risk factors and their associations with episode(s) of bac- teremia Variable Table 3. Previous history of infection before the current catheter and its association with episode(s) of bacteremia Episodes of bacteremia absence Men 1,158 Side of implantation (right) 1,844 Site of catheter implantation Femoral 11 Internal jugular 1,936 Subclavian 70 Hypertension 906 Immunodepression 401 Atherosclerosis 492 Obesity 127 Diabetes mellitus 243 Duration of catheter use >90 days 896 Age <50.3 years 513 ≥50.3–<63.6 years 526 ≥63.6–<72.6 years 519 ≥72.6 years 458 Wright-Khan index 0: low risk 1,259 1: medium risk 537 2: high risk 199 OR (95% CI) p value Type of infection presence 102 0.82 (0.61–1.11) NS 169 0.69 (0.43–1.09) NS 6 180 6 115 42 69 16 53 5.80 (2.12–15.88) 1 0.90 (0.39–2.13) 1.86 (1.38–2.52) 1.12 (0.79–1.61) 1.77 (1.30–2.42) 1.40 (0.81–2.41) 2.83 (2.01–4.00) <0.0001 NS <0.0001 NS <0.0001 123 2.19 (1.61–2.98) <0.0001 61 42 41 48 1 0.67 (0.44–1.01) 0.67 (0.44–1.01) 0.89 (0.60–1.33) 91 57 44 1 1.47 (1.04–2.08) 3.06 (2.07–4.52) <0.0001 NS <0.0001 diabetes mellitus (OR = 2.83; 95% CI = 2.01–4.00; p ! 0.0001), site of catheter implantation (femoral vs. jugular OR = 5.80; 95% CI = 2.12–15.88, and subclavian vs. jugular OR = 0.90; 95% CI = 0.39–2.13; p ! 0.0001), duration of catheter use (2.19; 95% CI = 1.61–2.98; p ! 0.0001), Wright-Khan comorbidity index (Wright 1 vs. 0: OR = 1.47; 95% CI = 1.04–2.08, and Wright 2 vs. 0: OR = 3.06; 95% CI = 2.07–4.52; p ! 0.0001) and previous history of CRB (OR = 3.07; 95% CI = 1.81–5.23; p ! 0.0001) were significant risk factors associated with the presence of bacteremia episode(s). By contrast, there were no significant differences between the presence and absence of bacteremia concerning gender, side of catheter implantation, immunodepression, age and obesity (fig. 4; tables 2, 3). Results of the multivariate analysis presented in table 4 revealed that previous history of a bacteremic episode (OR = 2.70; 95% CI = 1.56–4.68), diabetes mellitus (OR = 2.37; 95% CI = 1.65–3.39), duration of catheter use 190 days (OR = 1.85; 95% CI = 1.35–2.55) and hypertenRisk Factors for Catheter-Related Bacteremia 0 1 2 Episodes of bacteremia OR (95 % CI) absence presence 1,805 74 155 158 20 18 p value 1 <0.0001 3.15 (1.85-5.37) 1.30 (0.77-2.21) 0 = No previous history of infection; 1 = previous unrelated catheter infection; 2 = catheter-related bacteremic episode. Table 4. Multivariate analysis to assess the relative importance of the different risk factors associated with the presence of bacteremia episodes Variables OR 95% CI Previous history of bacteremia Diabetes mellitus Catheter use >90 days Hypertension 2.70 2.37 1.85 1.49 1.56–4.68 1.65–3.39 1.35–2.55 1.08–2.04 sion (OR = 1.49; 95% CI = 1.08–2.04) were still significant risk factors associated with the presence of bacteremia episodes. By contrast, no association between the presence of bacteremia and age, Wright-Khan index or atherosclerosis was reported. Survival Analysis As duration of catheter use 190 days was retained as a risk factor for CRB, a survival analysis of catheters with duration use 190 days (n = 1,054) has been performed to study the relationship between time and DCath survival. Figure 5 clearly demonstrated a relationship between time and DCath survival regarding catheters with duration use 190 days (n = 1,054) including both infection-free and salvaged ones after antimicrobial therapy. A comparison of patients who experienced bacteremia with those who did not showed no significant difference regarding survival (log-rank test: p = 0.45; fig. 6). Blood Purif 2009;28:21–28 25 1.00 DCath survival 0.75 Fig. 5. Survey study using the Kaplan-Meier method to study the relationship between time and bacteremic episode occurrence with catheters used 190 days (n = 1,054). This survey study clearly demonstrated a linear relationship between time and bacteremic episode occurrence thus predicting an episode of bacteremia in 50% of catheters after 4 years of use. Fig. 6. Comparison of DCath survival after 90 days of use in patients with or without bacteremia. a Overall survival of DCath (n = 1,054). b Survival comparison of DCath in patients experiencing one bacteremia episode or remaining free of infection. As shown, one episode of bacteremia effectively treated did not affect long-term DCath survival. 0.25 Product-limit estimate curve Censored observations 0 0 500 1,000 1,500 2,000 2,500 3,000 3,500 Time (days) Color version available online 5 1.00 Survival distribution function 1.00 Survival distribution function 0.50 0.75 0.50 0.25 Product-limit estimate curve Censored observations 0 0.75 No infection 0.50 infection inf = 0 Censored inf = 0 inf = 1 Censored inf = 1 0.25 0 0 a 20 40 Survival 60 80 100 0 b 20 40 Survival 60 80 100 6 Discussion The results of this 23-year prospective cohort study firstly reported an overall incidence of bacteremia of 0.514 per 1,000 catheter days. Among microorganisms involved in these bacteremia infections, S. aureus, coagulase-negative staphylococci, Enterobacter spp. and P. aeruginosa were mainly concerned. Hypertension, atherosclerosis, diabetes mellitus, previous history of CRB, site of catheter implantation, duration of catheter use and Wright-Khan index were the most important factors associated with the presence of bacteremia episode(s). 26 Blood Purif 2009;28:21–28 Our incidence of 0.514/1,000 catheter days (corresponding to 1,54 episodes/100 patient months) is one of the lowest incidence rates reported in the literature for catheters used as a permanent vascular access [13]. The incidence of catheter-related bacteremia is usually reported in the range 0.5–9/1,000 catheter days. It is worth noting that these results were achieved with uncuffed tunnelled catheters. This observation tends to prove that tunnelling and tight subcutaneous anchoring is a determining factor in preventing catheter-related infection [14] that competes with the antibiotic lock technique [15, 16]. Our data tend to superimpose EPIBACDIAL results in which 0.93 episodes/100 patient months were observed [5] in 988 Lemaire et al. patients from whom only 58 had catheters as permanent vascular access. In this study, the authors showed that vascular access (catheter vs. arteriovenous fistula) was a risk factor for the presence of bacteremia. These findings were later confirmed by the American prospective descriptive analysis performed by Stevenson et al. [17]. According to their study, the low incidence of bacteremia episodes observed in our study is all the more encouraging because we only studied patients with catheters as permanent vascular access. Such a low rate of bacteremia episodes is most likely linked to the hygienic and aseptic rules for catheter handling established in our center. However, of note, our study may have underestimated the number of infections due to the fact that bacteremic episodes may have been treated without catheter withdrawal. Regarding the nature of microorganisms involved in these bacteremia infections, our study is in total agreement with literature data [18]. Indeed, S. aureus represents the major causative pathogen (48%) involved in CRB infections [4, 5, 19], followed by coagulase-negative staphylococci [5, 20, 21] and Escherichia coli [22]. The high prevalence of S. aureus is in accordance with previous data in this clinical setting, a relationship that may be explained by the fact that HD patients may be chronic carriers of this agent. In agreement with several studies, 50–60% of HD patients are carriers of this germ [23–25], which confers an increased risk (!330) to develop bacteremia infections to this population [26]. In our study, a previous history of CRB was the major risk factors for bacteremia. This risk factor has already been reported in the literature and may be related to S. aureus nasal carriage [5]. Unfortunately, prospective data on nasal carriage were not available. However, the only observation we could make was that recurrence of bacteremia was not necessarily associated with previously occurring bacteria episodes, suggesting that a chronic carriage could not be incriminated. We identified the duration of catheter use 190 days as another risk factor for bacteremia. A few studies have reported this relationship between CRB and duration of catheter use [27, 28]. In their analysis of the cumulative hazard of developing CRB, Kairaitis and Gottlieb [27] reported that the instantaneous risk of developing bacteremia at any point in time was relatively constant, suggesting that catheter colonization and subsequent infection was a random event and that a ‘threshold’ duration at which the probability of bacteremia sharply increases did not exist [24]. This fact was not observed by Hung et al. [28] who reported a 50% risk of bacteremia after 2 months of catheterization [25]. In agreement with their study, we also found a ‘threshold’ duration, but only after 90 days. Such a difference may be explained by the lower number of catheters studied and the shorter follow-up period in the study by Hung et al. [28]. In our study, diabetes was also reported as a risk factor for bacteremia. Controversy exists regarding diabetes as a risk factor for bacteremia in the literature. Indeed, it has been largely reported as a risk factor in US studies [17, 29, 30] whereas no European study [4, 5, 19] could demonstrate it. This discordance may be explained by an important difference in diabetes prevalence between Europe (18% in France for example) and the USA (50%). Even though in our study the prevalence of diabetes was even lower than that of France (14.6%), the number of patients was twice the number of patients included in EPIBACDIAL, which should confer a more important statistical power to our study. Diabetes as a risk factor for bacteremia is all the more foreseeable because diabetic patients are more often confronted with S. aureus bacteremia, a risk which is clearly associated with a more frequent nasal carriage [24, 31]. Hypertension was reported as a risk factor for bacteremia, but with a low statistical power maybe due to the increased prevalence of this comorbidity in our population (46.1%). Femoral site of catheter implantation was also considered as a high risk factor for bacteremia compared to subclavian and internal jugular implantation sites, which is in total agreement with previous studies [32]. One can speculate that this could result from the proximity of areas with precarious hygiene even though such catheters are tunneled. In conclusion, reducing CRB is still a challenge for nephrologists in order to decrease patient morbidity and mortality. Considering our results, in order to reduce the incidence of CRB, special attention should be given to patients with a previous history of bacteremia, patients with diabetes, i.e. carriers of S. aureus should receive systematic disinfection by nasal and armpit application of mupirocin [33, 34]. In addition, catheter handling should also be subjected to strict hygienic and aseptic rules concomitant with routine prophylactic use of antiseptic lock solutions. Risk Factors for Catheter-Related Bacteremia Blood Purif 2009;28:21–28 Acknowledgments The authors would like to thank the nephrologists from the following dialysis centers: AIDER, Montpellier; Centre d’Hémodialyse Languedoc Méditerranéen, Montpellier, and Clinique Les Genêts, Narbonne, for their collaboration. 27 References 1 US Renal Data System: USRDS 2005 Annual Data Report. Bethesda, National Institutes of Health, Department of Health and Human Services, 2005. 2 Pastan S, Soucie JM, McClellan WM: Vascular access and increased risk of death among hemodialysis patients. Kidney Int 2002; 62: 620–626. 3 Besarab A: Vascular access: issues and management. Contrib Nephrol 2004;142:29–46. 4 Bonello M, Levin NW, Ronco C: History and evolution of the vascular access for hemodialysis. Contrib Nephrol 2004; 142:1–13. 5 Konner K: Complications of the vascular access for hemodialysis. Contrib Nephrol 2004;142:193–215. 6 Rayner HC, Besarab A, Brown WW, et al: Vascular access results from the Dialysis Outcomes and Practice Patterns Study (DOPPS): performance against Kidney Disease Outcomes Quality Initiative (K/DOQI) Clinical Practice Guidelines. Am J Kidney Dis 2004;44(5 suppl 2):22–26. 7 Hoen B, Kessler M, Hestin D, Mayeux D: Risk factors for bacterial infections in chronic haemodialysis adult patients: a multicentre prospective survey. Nephrol Dial Transplant 1995;10:377–381. 8 Hoen B, Paul-Dauphin A, Hestin D, Kessler M: EPIBACDIAL: a multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients. J Am Soc Nephrol 1998;9:869–876. 9 Jean G, Charra B, Chazot C, et al: Risk factor analysis for long-term tunneled dialysis catheter-related bacteremias. Nephron 2002; 91:399–405. 10 Canaud B, Beraud JJ, Joyeux H, Mion C: Internal jugular vein cannulation with two silicone rubber catheters: a new and safe temporary vascular access for hemodialysis. Thirty months’ experience. Artif Organs 1986;10:397–403. 11 Wright LF: Survival in patients with endstage renal disease. Am J Kidney Dis 1991;17: 25–28. 12 Khan IH, Catto GR, Edward N, et al: Influence of coexisting disease on survival on renal-replacement therapy. Lancet 1993; 341: 415–418. 13 Mokrzycki MH, Zhang M, Cohen H, Golestaneh L, Laut JM, Rosenberg SO: Tunnelled haemodialysis catheter bacteraemia: risk factors for bacteraemia recurrence, infectious complications and mortality. Nephrol Dial Transplant 2006;21:1024–1031. 28 14 Aslam N, Lambie M, Ayub V, Stein A: Use of tunnelled haemodialysis catheters at the start of haemodialysis – success rates and definition of infection. Nephrol Dial Transplant 2007;22:1799–1800. 15 Kim SH, Song KI, Chang JW, Kim SB, Sung SA, Jo SK, Cho WY, Kim HK: Prevention of uncuffed hemodialysis catheter-related bacteremia using an antibiotic lock technique: a prospective, randomized clinical trial. Kidney Int 2006;69:161–164. 16 Nori US, Manoharan A, Yee J, Besarab A: Comparison of low-dose gentamicin with minocycline as catheter lock solutions in the prevention of catheter-related bacteremia. Am J Kidney Dis 2006;48:596–605. 17 Stevenson KB, Adcox MJ, Mallea MC, et al: Standardized surveillance of hemodialysis vascular access infections: 18-month experience at an outpatient, multifacility hemodialysis center. Infect Control Hosp Epidemiol 2000;21:200–203. 18 Taylor G, Gravel D, Johnston L, Embil J, Holton D, Paton S; Canadian Hospital Epidemiology Committee; Canadian Nosocomial Infection Surveillance Program: Prospective surveillance for primary bloodstream infections occurring in Canadian hemodialysis units. Infect Control Hosp Epidemiol 2002;23:716–720. 19 Jean G, Charra B, Chazot C, Vanel T, Terrat JC, Hurot JM: Long-term outcome of permanent hemodialysis catheters: a controlled study. Blood Purif 2001;19:401–407. 20 Foley RN, Guo H, Snyder JJ, et al: Septicemia in the United States dialysis population, 1991 to 1999. J Am Soc Nephrol 2004; 15: 1038– 1045. 21 Marr KA, Sexton DJ, Conlon PJ, et al: Catheter-related bacteremia and outcome of attempted catheter salvage in patients undergoing hemodialysis. Ann Intern Med 1997; 127:275–280. 22 Hajjar J, Girard R, Marc JM, et al: Surveillance of infections in chronic hemodialysis patients (in French). Nephrologie 2004; 25: 133–140. Blood Purif 2009;28:21–28 23 Kirmani N, Tuazon CU, Murray HW, et al: Staphylococcus aureus carriage rate of patients receiving long-term hemodialysis. Arch Intern Med 1978;138:1657–1659. 24 Goldblum SE, Ulrich JA, Goldman RS, Reed WP: Nasal and cutaneous flora among hemodialysis patients and personnel: quantitative and qualitative characterization and patterns of staphylococcal carriage. Am J Kidney Dis 1982;2:281–286. 25 Yu VL, Goetz A, Wagener M, et al: Staphylococcus aureus nasal carriage and infection in patients on hemodialysis. Efficacy of antibiotic prophylaxis. N Engl J Med 1986; 315:91– 96. 26 Nielsen J, Kolmos HJ, Espersen F: Staphylococcus aureus bacteraemia among patients undergoing dialysis – focus on dialysis catheter-related cases. Nephrol Dial Transplant 1998;13:139–145. 27 Kairaitis LK, Gottlieb T: Outcome and complications of temporary haemodialysis catheters. Nephrol Dial Transplant 1999; 14: 1710–1714. 28 Hung KY, Tsai TJ, Yen CJ, Yen TS: Infection associated with double lumen catheterization for temporary haemodialysis: experience of 168 cases. Nephrol Dial Transplant 1995;10:247–251. 29 Powe NR, Jaar B, Furth SL, et al: Septicemia in dialysis patients: incidence, risk factors, and prognosis. Kidney Int 1999; 55: 1081– 1090. 30 Allon M, Depner TA, Radeva M, et al: Impact of dialysis dose and membrane on infectionrelated hospitalization and death: results of the HEMO Study. J Am Soc Nephrol 2003;14: 1863–1870. 31 Chi CY, Wong WW, Fung CP, et al: Epidemiology of community-acquired Staphylococcus aureus bacteremia. J Microbiol Immunol Infect 2004;37:16–23. 32 Goetz AM, Wagener MM, Miller JM, Muder RR: Risk of infection due to central venous catheters: effect of site of placement and catheter type. Infect Control Hosp Epidemiol 1998;19:842–845. 33 Kluytmans JA, Manders MJ, van Bommel E, Verbrugh H: Elimination of nasal carriage of Staphylococcus aureus in hemodialysis patients. Infect Control Hosp Epidemiol 1996; 17:793–797. 34 Davey P: Eradication of nasal carriage of Staphylococcus aureus: is it cost-effective? J Hosp Infect 1998;40(suppl B):S31–S37. Lemaire et al.