- Email: [email protected]
A Randomized Double-Blind Controlled Trial of Taurolidine-Citrate Catheter Locks for the Prevention of Bacteremia in Patients Treated With Hemodialysis
Dialysis A Randomized Double-Blind Controlled Trial of Taurolidine-Citrate Catheter Locks for the Prevention of Bacteremia in Patients Treated With Hemodialysis Laurie R. Solomon, FRCP,1 John S. Cheesbrough, FRCPath,2 Leonard Ebah, MRCP,1 Tamer Al-Sayed, MRCP,1 Michael Heap, RGN,1 Nick Millband, MBCS,1 Dee Waterhouse, RGN,3 Sandip Mitra, FRCP,3 Alan Curry, PhD,4 Rema Saxena, MRCP,5 Rammohan Bhat, MRCP,5 Michael Schulz, FRCP,5 and Peter Diggle, PhD6 Background: Bacteremia is a major cause of morbidity in patients using intravascular catheters. Interdialytic locking with antibiotics decreases the incidence of bacteremia, but risks antibiotic resistance. Taurolidine is a nontoxic broad-spectrum antimicrobial agent that has not been associated with resistance. Preliminary evidence suggests that taurolidine-citrate locks decrease bacteremia, but cause flow problems in established catheters. Study Design: Double-blind randomized controlled trial. Intervention: Interdialytic locking with taurolidine and citrate (1.35% taurolidine and 4% citrate) compared with heparin (5,000 U/mL) started at catheter insertion. Setting & Participants: 110 adult hemodialysis patients with tunneled cuffed intravascular catheters inserted at 3 centers in Northwest England. Outcomes & Measurements: Primary end points were time to first bacteremia episode from any cause and time to first use of thrombolytic therapy. Results: There were 11 bacteremic episodes in the taurolidine-citrate group and 23 in the heparin group (1.4 and 2.4 episodes/1,000 patient-days, respectively; P ⫽ 0.1). There was no significant benefit of taurolidine-citrate versus heparin for time to first bacteremia (hazard ratio, 0.66; 95% CI, 0.2-1.6: P ⫽ 0.4). Taurolidine-citrate was associated with fewer infections caused by Gram-negative organisms than heparin (0.2 vs 1.1 infections/1,000 patient-days; P ⫽ 0.02); however, there was no difference for Gram-positive organisms (1.1 vs 1.2 infections/1,000 patient-days; P ⫽ 0.8). There was a greater need for thrombolytic therapy in the taurolidine-citrate versus heparin group (hazard ratio, 2.5; 95% CI, 1.3-5.2; P ⫽ 0.008). Limitations: Small sample size. The study included bacteremia from all causes and was not specific for catheter-related bacteremia. Conclusions: Taurolidine-citrate use did not decrease all-cause bacteremia and was associated with a greater need for thrombolytic treatment. There was a decrease in infections caused by Gram-negative organisms and a trend to a lower frequency of bacteremia, which warrants further study. Am J Kidney Dis 55:1060-1068. © 2010 by the National Kidney Foundation, Inc. INDEX WORDS: Bacteremia; catheter; hemodialysis; lock solution; randomized trial; taurolidine.
I
nfection is a major cause of morbidity and mortality in hemodialysis patients.1 Although decreasing the number of intravascular catheters and increasing the proportion of arteriovenous fistulas decrease the incidence of infection, catheters cannot be eliminated entirely. Pre-
vention of bacteremia is crucial. One approach is to use catheter locks containing an antibiotic. Decreased bacteremia has been reported using gentamicin, cefotaxime, and minocycline and confirmed using meta-analysis.2-6 Risks include toxicity from leakage of the antibiotic into the
From the 1Renal Unit and 2Department of Microbiology, Lancashire Teaching Hospitals, Royal Preston Hospital, Lancashire; 3Renal Unit and 4Health Protection Agency, Manchester Royal Infirmary, Manchester; 5Renal Unit, Royal Liverpool Hospitals, Liverpool; and 6CHICAS, Division of Medicine, School of Health and Medicine, University of Lancaster, Lancaster, UK. Received May 23, 2009. Accepted in revised form November 27, 2009. Originally published online as doi:10.1053/ j.ajkd.2009.11.025 on March 8, 2010.
Trial registration: www.isrctn.org; study number: ISRCTN51902376. Address correspondence to Laurie R. Solomon, FRCP, Renal Unit, Lancashire Teaching Hospitals, Royal Preston Hospital, Sharoe Green Lane, Fulwood, Preston, Lancashire PR2 9HT, UK. Email: [email protected] © 2010 by the National Kidney Foundation, Inc. 0272-6386/10/5506-0013$36.00/0 doi:10.1053/j.ajkd.2009.11.025
1060
American Journal of Kidney Diseases, Vol 55, No 6 (June), 2010: pp 1060-1068
Taurolidine-Citrate Catheter Locks
circulation and the development of antibiotic resistance.3,7,8 An alternative is taurolidine-citrate. Taurolidine (2H-1,2,4-thiadiazine-4,4=-methylenebis[tetrahydro-1,1,1=,1=-tetraoxide]) yields methylol tauranimide components that bind to bacterial and fungal cell walls, causing irreversible damage. This confers very broad-spectrum antimicrobial activity and decreases the development of biofilm.9 Resistance has not been reported and is unlikely because the mode of action resembles a disinfectant, rather than an antibiotic, which exploits a single metabolic pathway.10 Uncontrolled studies suggest that taurolidine locks result in a very low bacteremia rate in hemodialysis and nutrition catheters,11-13 decrease infection in central venous access in children,14 and eradicate catheter-related bloodstream infections unresponsive to systemic antibiotic therapy.15 One nonrandomized controlled trial observed higher bacteremia-free survival at 90 days with taurolidine-citrate than with heparin, but a greater requirement for thrombolytics to maintain catheter patency. This was undertaken in established catheters, which may have contained bacterial biofilm or intraluminal deposits of blood component origin.16 Increased thrombosis might not have occurred if taurolidinecitrate treatment had been started at catheter insertion. Another open-label study randomly assigned patients with temporary and tunneled catheters to taurolidine-citrate or heparin and observed no catheter-related sepsis with taurolidine-citrate compared with 4 episodes with heparin. This study included few tunneled catheters and no information about patency.17 There has been no trial testing taurolidine-citrate from the time of catheter insertion. We undertook a randomized double-blind trial comparing taurolidine-citrate catheter locks with heparin catheter locks started at the time of catheter insertion in hemodialysis patients using tunneled cuffed intravascular catheters. The objective was to determine the incidence of bacteremia and determine whether taurolidine-citrate use is associated with a greater need for thrombolysis.
METHODS Patients The study was undertaken in Northwest England. Catheters were placed at 3 centers and patients were dialyzed at
1061 13 centers. Recruitment took place between November 2006 and May 2008. Adult patients aged ⬎18 years receiving tunneled intravascular catheters for hemodialysis and able to give informed consent were eligible. The Cumbria and Lancashire B Research Ethics committee approved the study (06/Q1309/30), and all patients gave written consent.
Study Procedures Independent pharmacists undertook randomization on a 1:1 basis using computer-generated randomized permuted blocks of 10 patients, which were stratified among the 3 main centers. Physicians responsible for the patient’s care undertook recruitment and obtained consent. Ampoules were labeled “trial catheter lock” and with the patient’s name. All study personnel and participants were blinded to treatment assignment throughout the study until the database was complete. Patients were randomly assigned to receive either taurolidine-citrate or unfractionated heparin (5,000 U/mL) at catheter insertion and after every dialysis treatment until a trial end point. The preferred site of insertion was the right followed by the left internal jugular vein. Right-sided catheters were inserted in an operating room or purpose-built treatment room by a small number of skilled operators. Leftand some right-sided catheters were placed using fluoroscopic guidance in the Radiology Department. Antibiotic prophylaxis was not given. Routine antimicrobial body washes were introduced during the course of the study. Taurolidine-citrate (Taurolock; a solution of 1.35% taurolidine and 4% citrate; Tauropharm AG, www.tauropharm.de) and heparin (porcine heparin sodium [European Pharmacopoeia] at 5,000 U/mL with 1% benzylic acid as preservative [B. Braun AG, www.bbraun.com]) were supplied in identical ampoules. Nurses drew trial solution into a syringe after dialysis and injected the exact locking volume into the catheter lumens. Compliance was checked every 2 weeks. If there was a lapse in the use of trial locks, a note was made and the patient continued in the trial using the trial solution. Exit sites were inspected after every dialysis treatment, cleaned weekly with 2% chlorhexidine in 70% isopropyl alcohol, and covered with a see-through semipermeable membrane (IV 3000 1-HAND; Smith and Nephew, www. smith-nephew.com). If a catheter was replaced, patients could continue in the same arm of the study. Reasons for discontinuation were catheter removal for any reason, kidney transplant, conversion to peritoneal dialysis therapy, transfer to a nonparticipating dialysis unit, development of heparin-induced thrombocytopenia, death, or withdrawal from the study by the patient or attending physician.
Diagnosis and Treatment of Bacteremia The decision to obtain blood cultures was based on symptoms of infection, such as fever (temperature ⬎37.5°C) or rigors associated with dialysis. A single positive blood culture bottle defined a bacteremic episode. Blood culture samples were obtained using aseptic technique from the arterial port of the dialysis circuit. If the same organism was isolated in cultures ⬍3 weeks apart, it was considered the same infection. When different organisms were identified or
1062 if cultures were ⬎3 weeks apart, these were considered separate episodes. Vancomycin and gentamicin (either individually or together) were administered intravenously and as catheter locks for suspected catheter infections according to established protocols for either 2 or 4 weeks, depending on the organism. Oral flucloxacillin or rifampicin was added for Staphylococcus aureus. Doses were adjusted according to blood concentrations. Antibiotic catheter locks were made up in heparin (final concentration, 5,000 U/mL). Catheter removal was considered if the patient remained unwell for ⬎72 hours after starting antibiotic therapy. Exit-site infection was diagnosed on the basis of discharge, erythema, swelling, tenderness, and isolation of a pathogenic organism from a swab.
Thrombolytic Therapy Thrombolytic therapy was given when dialysis was impossible because of poor or no flow. A locking volume of either urokinase (5,000 U/mL; prepared from urokinase powder supplied as 25,000-unit vials; Syner-KINASE; Syner-Med, www.syner-med.com) or tissue plasminogen activator (1 mg/mL; reconstituted from powder supplied in 10-mg vials; Alteplase; Boehringer Ingelheim, www.boehringer-ingelheim. com) was instilled into both catheter lumens for half an hour before retrying. If flow remained poor, a locking dose of thrombolytic was left in the catheter throughout the next interdialytic period. Three doses could be given after 1 episode before considering other options, such as catheter removal. No pumping was used. After successful thrombolysis, attending physicians could add 0.5 mL of 5,000-U/mL heparin to each 5 mL of trial lock solution. This gave a final concentration of 455 U/mL of heparin in the taurolidine-citrate arm, but left the heparin arm unchanged at 5,000 U/mL.
Electron Microscopy of Removed Catheters Two catheter segments (from the tunnel and tip) were cut aseptically and fixed in formalin and osmium tetroxide. After dehydration, catheter sections were immersed in propylene oxide, which caused the catheter material to swell and adherent material to detach. Detached tissue was embedded in resin. Ultrathin sections were cut and stained before examination under a Philips CM10 electron microscope (FEI, www.fei.com).
Solomon et al yes/no), analysis was based on time to first bacteremia event to obtain a more efficient analysis. We report P values and confidence intervals (CIs) for hazard ratios.
Primary and Secondary Outcomes The primary outcome was time to first bacteremia event. Secondary analyses were undertaken for the total number of bacteremic episodes and Gram-positive and Gram-negative infections.
Statistical Analysis A Cox proportional hazards model was used for time to first bacteremia and time to first use of thrombolytic therapy, in each case estimating and testing the treatment effect after stratifying by the 3 main centers and adjusting for age. When calculating catheter survival, recovery of kidney function, kidney transplant, conversion to peritoneal dialysis therapy, availability of alternative vascular access, and transfer to another dialysis unit were considered favorable outcomes. Infection, occlusion, death, and patient or physician choice to leave the study were considered unfavorable. Fisher exact test was used for binary (yes/no) outcomes, and t test, for continuous outcomes, but implemented using Monte Carlo in the case of dialysis vintage because of its highly nonnormal distribution. Counted outcomes were analyzed as rates of occurrence/1,000 patient-days by fitting a Poisson log-linear model with the logarithm of patient-days as an offset. P ⬍ 0.05 is considered significant.
RESULTS Of 256 patients who had tunneled catheters placed during the period of the trial, 110 were recruited (55 taurolidine-citrate, 55 heparin; Fig 1). Most exclusions were logistical given the short time between the decision to insert a catheter and placement or because the patient lacked capacity to consent. At center C, 2 patients were
Sample Size Preliminary data from center A showed that bacteremia occurred in ⬃50% of patients with catheters after 12 months. The all-cause bacteremia rate was 3.3 events/1,000 patientdays for all organisms. We calculated that 31 patients in each group would be needed to show a decrease in the incidence of bacteremia from 50% in the heparin group to 10% in the taurolidine group with a power of 0.9 at the 0.05% level of significance (Fisher exact test). It was believed that allowance should be made for bias of published results and improvements in care during the course of the study. It was decided to include 50 patients in each group and continue 6 months after recruiting the last patient. Although the study was powered on the basis of a binary outcome (bacteremia,
Figure 1. Trial flow diagram (intention to treat analysis). Table 4 lists details of follow-up.
Taurolidine-Citrate Catheter Locks
1063
Table 1. Demographic Details by Type of Catheter Lock TaurolidineCitrate
Heparin
No. of patients
53
54
No. of catheters
56
58
Age (y)
P
0.9
59.8 ⫾ 14.7 56.7 ⫾ 17.4 0.3
Men
26
41
0.009
Race White Asian Black Mixed/other
47 4 1 1
49 5 0 0
0.5a
0 (0-13)
0 (0-6)
0.4
5 7 5 6
3 5 7 11
30
28
Dialysis vintage (y) Cause of ESRD Cystic kidney disease Diabetic nephropathy Glomerulonephritis Hypertensive/ ischemic Other
Note: Values expressed as mean ⫾ standard deviation, median (range), or number. Abbreviation: ESRD, end-stage renal disease. a P is for white versus other.
excluded because the responsible physician considered that they were at high risk of infection because of previous infections and that they lived in an unhygienic environment. Baseline characteristics are listed in Table 1. At the 3 centers, 70, 20, and 20 patients were recruited, representing 62%, 33%, and 33% of catheters placed during the study period, respectively. There was no significant difference between the 2 groups in catheter type (P ⫽ 0.8) or location (P ⫽ 0.9; Table 2). Three patients, 2 in the taurolidine-citrate group and 1 in the heparin group, were allocated trial numbers, but never needed dialysis. Thus, an intention-totreat analysis was performed on 53 patients in the taurolidine-citrate group and 54 in the heparin group. We estimated that trial locks were administered ⬎95% of the time. One patient in the taurolidine-citrate arm had a positive blood culture result for Aerococcus urinate, but was documented before unblinding to have lapsed trial locks for 3 weeks beforehand. Lapse of protocol resulted in default use of heparin locks. This patient was included in the taurolidine-citrate
arm of the intention-to-treat analysis and the heparin arm of the as-treated analysis. Seven patients (3 taurolidine-citrate, 4 heparin) had 2 catheters. Reasons for removal of the first catheter were occlusion (1 heparin, 3 taurolidinecitrate), catheter fell out (1 heparin), infection (1 heparin), and exit-site infection (1 heparin). There were 17,771 catheter-days (8,129 taurolidine-citrate, 9,642 heparin). There were significantly more men in the heparin group despite randomization, but other parameters were similar between the 2 groups (Table 1). Time to First Bacteremia Intention-to-treat analysis showed no significant difference between taurolidine-citrate and heparin with respect to time to first bacteremia after censoring for favorable outcomes (Fig 2). The estimated hazard ratio in favor of taurolidinecitrate was 0.66 (95% CI, 0.28-1.59; P ⫽ 0.4). Age, sex and race were not significant (P ⫽ 0.4, 0.1 and 0.5, respectively). Years on dialysis therapy was borderline significant (P ⫽ 0.05). The secondary as-treated analysis showed no significant difference for taurolidine-citrate over heparin with respect to time to first bacteremia. The estimated hazard ratio in favor of taurolidinecitrate was 0.55 (95% CI, 0.23-1.35; P ⫽ 0.2). Table 2. Catheter Type and Location by Type of Catheter Lock Taurolidine-Citrate
Heparin
53
54
56
58
First catheter type Ash Split Cathb Bard HemoSplitc Circle-Cd Bard HemoGlidec Quinton-Permcathe Bio-Flex Tesiob
31 1 2 5 3 14
34 1 2 5 1 15
Location Internal jugular Subclavian
55 1
57 1
No. of patients No. of catheters a
a
All were cuffed catheters. Medcomp: www.medcompnet.com. c Bard: www.barduk.com. d Horizon Medical Products: www.opitsourcebook.com. e Kendall: www.kendallhq.com. b
1064
Solomon et al
and 6 in the heparin group (P ⫽ 0.9). In the taurolidine-citrate group, all were S aureus. In the heparin group, 4 were S aureus and 2 were coliforms. Four of the 13 infections were associated with positive blood cultures for the same organism (2 heparin, 2 taurolidine-citrate).
Figure 2. Time to first bacteremia. Patients were censored for catheter removal because of recovery of kidney function, kidney transplant, conversion to peritoneal dialysis therapy, availability of alternative vascular access, irreversible occlusion, exit-site infection (without bacteremia), or death or for withdrawal from the trial as a result of patient or physician choice, transfer to a nonparticipating dialysis unit, or development of heparin-induced thrombocytopenia. Abbreviation: TC, taurolidine-citrate.
Bacteremic Episodes There were 11 bacteremic episodes in 9 patients in the taurolidine-citrate group compared with 23 episodes in 16 patients in the heparin group (Table 3). In the heparin group, 3 patients experienced 2 bacteremic episodes and 2 patients experienced 3 bacteremic episodes. In the taurolidine-citrate group, 2 patients experienced 2 bacteremic episodes. The second and third organisms were different species in 6 patients. Blood cultures from 1 patient in the heparin group grew a coagulase-negative staphylococcus with the same antibiogram after a 6-week interval. Nine different species were isolated from patients using heparin compared with 3 different species from patients using taurolidine-citrate. There were 11 and 2 Gram-negative infections in the heparin and taurolidine-citrate groups, respectively. The infection rate for Gram-negative organisms was significantly lower in the taurolidinecitrate group compared with the heparin group (P ⫽ 0.02), but there was no difference in the number of Gram-positive infections (P ⫽ 0.8; Table 3). Exit Sites Pathogenic organisms were cultured from exit sites on 13 occasions. There were 7 positive culture results in the taurolidine-citrate group
Use of Thrombolytic Therapy More patients in the taurolidine-citrate group needed thromobolytic therapy at least once (taurolidine-citrate, 28 of 53; heparin, 14 of 54; P ⫽ 0.006). There was a significant adverse effect of taurolidine-citrate with respect to time to first use of Table 3. Number of Bacteremic Episodes and Organisms Isolated
No. of patients a
Bacteremic episodes Gram positivea,b Staphylococcus aureus (MSSA) Staphylococcus aureus (MRSA) Other Gram positive Gram negativea,d No. of infections/patient 1 2 3 No. of catheter-days
TaurolidineCitrate
Heparin
53
54
11 (1.4) 9 (1.1) 5
23 (2.4) 12 (1.2) 6
1c
2
3 2 (0.2)
P
0.1 0.8
4 11c (1.1) 0.02 0.3
7 2 0
11 3 2
8,129
9,642
Note: Based on intention-to-treat analysis. Abbreviations: MRSA, methicillin-resistant S aureus; MSSA, methicillin-sensitive S aureus. a Figures outside brackets are actual numbers. Figures in brackets are per 1,000 catheter-days. b Staphylococcus aureus (taurolidine citrate, 6; heparin, 8), coagulase-negative Staphylococcus (taurolidine citrate, 2; heparin, 4), Aerococcus urinae (taurolidine citrate, 1). c On 2 occasions, 2 organisms were grown from the same blood culture. One patient in the heparin group had Klebsiella and Enterobacter asbiriae; 1 patient in the taurolidine-citrate group had methicillin-resistant S aureus and coagulase-negative staphylococcus from the same blood culture. The second organism is not included in this table. d Escherichia coli (taurolidine citrate, 2; heparin, 2), Klebsiella oxytoca (heparin, 1), Delftia acidovorans (heparin, 1), Enterobacter aerogenes (heparin, 1), Haemophilus parainfluenzae (heparin, 1), Serratia marcescens or S liquefaciens (heparin, 4), and Strenophonomonas maltophilia (heparin, 1).
Taurolidine-Citrate Catheter Locks
1065
infection (P ⫽ 0.8). Eight taurolidine-citrate and 3 heparin catheters were removed because of catheter occlusion (P ⫽ 0.06). There were 11 deaths in the taurolidine-citrate group compared with 8 deaths in the heparin group (P ⫽ 0.2). Differences Between Centers
Figure 3. Time to first use of thrombolytic therapy. Patients were censored for catheter removal because of recovery of kidney function, kidney transplant, conversion to peritoneal dialysis therapy, availability of alternative vascular access, bacteremia, exit-site infection, or death or for withdrawal from the trial as a result of patient or physician choice, transfer to a nonparticipating dialysis unit, or development of heparin induced thrombocytopenia. Abbreviation: TC, taurolidine-citrate.
thrombolytic therapy (Fig 3). The estimated hazard ratio against taurolidine-citrate was 2.5 (95% CI, 1.3-5.2; P ⫽ 0.008). In addition, 8 catheters in the taurolidinecitrate group and 3 in the heparin group were lost because of occlusion (P ⫽ 0.06). There were 79 administrations of thrombolytic to the taurolidinecitrate group and 38 to the heparin group. Thus, thrombolysis was successful ⬎90% of the time. Catheter Survival Median survival for first catheters was 271 days (95% CI, 245-297) for taurolidine-citrate and 358 days (95% CI, 270-445) for heparin after censoring for favorable outcomes, but including all deaths and withdrawals for patient or physician choice as adverse outcomes (P ⫽ 0.3). Reasons for Trial Termination Thirty-two taurolidine-citrate and 36 heparin catheters were removed for reasons unrelated to the trial, including alternative access, peritoneal dialysis therapy, recovery of kidney function, kidney transplant, transfer to another dialysis unit, and physician or patient choice (Table 4). Two taurolidine-citrate and 8 heparin catheters were removed because of persistent signs of infection after a positive blood culture (P ⫽ 0.1). Two taurolidine-citrate and 3 heparin catheters were removed because of exit-site or tunnel
Subgroup analysis was undertaken for center A, which had provided the background data for sample-size calculation, contributed most study participants, and recruited the greatest proportion of eligible patients, patient-days, and bacteremic episodes. Analysis was undertaken because of possible selection of lower risk patients from centers B and C, which had low rates of infection in the heparin arm (0 and 1.51/1,000 catheter-days). Center A contributed 7 bacteremic episodes in the taurolidine-citrate arm and 21 in the heparin arm (P ⫽ 0.02). Estimated hazard ratios for time to first bacteremia were 0.5 (95% CI, 0.2-1.2; P ⫽ 0.12) in the intention-totreat analysis and 0.37 (95% CI, 0.13-1.02; P ⫽ Table 4. Reasons for Catheter Removal or Trial Termination
No. of patients No. of catheters Bacteremia Exit-site infection Removal for occlusion/flow problems Patient choice Physician choice Heparin-induced thrombocytopenia Recovered renal function Alternative access available Conversion to peritoneal dialysis Transplant Transfer to another dialysis unit Reached end of trial Catheter fell out Deaths
TaurolidineCitrate
Heparin
53 56 2 2 8
54 58 8 3 3
0.1 0.8 0.1
2 1 1
2 0 0
0.8 0.2 0.2
4 17 3
4 15 1
0.7 0.3 0.2
1 3
2 4
0.7 0.9
1 0 11a
7 1 8b
0.05 0.3 0.2
P
a Cause of death unknown (6), cardiac (2), staphylococcus sepsis (1), cirrhosis (1), miliary tuberculosis (1), and myeloma (1). b Cause of death unknown (4), infected toe (1), mesothelioma (1), myeloma (1), and peripheral ischemia (1).
1066
0.06) in the as-treated analysis in favor of taurolidine-citrate. Intraluminal Material at Time of Removal Fourteen catheters (4 taurolidine-citrate, 10 heparin) were examined using transmission electron microscopy. All showed deposition of material derived from red blood cells, fibrin, and platelets at the tip. In the tunnel segment where the locking solution may be maintained without blood dilution between hemodialysis sessions, all except 4 catheters showed similar deposits. There were bacteria in 2 catheters: 1 in the heparin arm had an S aureus bacteremia, and organisms resembling Gram-positive cocci surrounded by polymorphs were seen in the tunnel and tip segments; in 1 catheter in the taurolidinecitrate arm, 1 Gram-negative rod was seen among red blood cells at the tip, but this was not associated with clinical or microbiological evidence of infection. There was no bacterial biofilm in any catheter.
DISCUSSION In this randomized double-blind controlled trial of taurolidine-citrate catheter locks in hemodialysis patients, we found no difference in time to first infection, but a significant decrease in the number of Gram-negative infections. There was a greater need for thrombolytic therapy in taurolidine-citrate locks compared with heparin locks. There was no significant difference in catheter survival. Results are consistent with previous open-label trials and observational studies.12,16,17 The distribution of Gram-negative organisms was skewed: both infections in the taurolidinecitrate arm were with typical endogenous flora, whereas 9 in the heparin arm were caused by organisms often associated with environmental reservoirs. This result suggests that there was protection against bacterial contamination during connect-disconnect procedures. However, there was no difference in the incidence of Grampositive organisms, including S aureus. In vitro studies indicate that the concentration of taurolidine was sufficient to kill S aureus.9,10 One possibility is that S aureus ingresses by a different portal, such as the exit site. We decided to include all positive blood cultures without attempting to differentiate catheter-related blood-
Solomon et al
stream infections from those originating at other foci because proving that bacteremia is catheter related requires simultaneous blood cultures from a peripheral vein and the catheter. We wanted to avoid unnecessary venipuncture and continue the standard practices of the participating units. A positive blood culture was a clear and verifiable measure. Some patients in both arms were very ill for other reasons. One of 2 Gramnegative infections in the taurolidine-citrate group occurred in a patient with cirrhosis and spontaneous bacterial peritonitis. Four episodes of bacteremia, 2 in each group, were associated with exit-site infection. The finding of approximately equal numbers of positive swabs from exit sites in the 2 groups is similar to a previous report17 and suggests that exit-site infections have a different pathogenesis. Bacterial infections related to dialysis catheters can develop through extra- and intraluminal routes.18 Catheter locks combat intraluminal contamination and cannot be expected to eliminate bacteremia originating in the exit site. This may require different measures. Mupirocin or polysporin cream applied to the exit site may be effective prophylaxis.19,20 We failed to show bacterial glycocalyx in any catheter. We found accumulation of fibrin, erythrocytes, and platelets at the tip of all catheters in both groups. Therefore, it was not possible to determine whether taurolidine-citrate destroys established bacterial biofilm. Bacterial biofilm may not be universal.21 This warrants further study. The relative decrease in bacteremia should be viewed in light of the incidence in the heparin arm because a low rate of infection in the control group reduces the scope for improvement. The relative decrease using taurolidine-citrate in our study was less than the 90% decrease in bacteremia reported by Allon,16 but the rate in the heparin arm in that study was 5.6 events/1,000 catheter-days (compared with 2.4 events/1,000 catheter-days in our trial). A higher incidence also was observed in the heparin arms of studies using gentamicin and 30% citrate.2,22 During the study period, the UK Department of Health encouraged rigorous infection control discipline to decrease methicillin-resistant S aureus infection.23 This may have contributed to a low incidence of bacteremia in our study.
Taurolidine-Citrate Catheter Locks
We performed subgroup analysis at center A, which suggested a significant decrease in the rate of infection and trend to decreased hazard ratio for time to first infection. Center A contributed most patients and catheters placed during the study period. It is possible that this was more representative of the general dialysis population and that the other 2 centers recruited lower risk patients. Furthermore, fewer taurolidine-citrate patients had ⬎1 bacteremic episode. Taurolidinecitrate therefore may be of greatest benefit in patients at high risk of infection. The main disadvantage of taurolidine-citrate was a greater need for thrombolytic therapy to maintain catheter patency. This was successful in most cases; however, catheter removal was required in 8 patients in the taurolidine-citrate group compared with 3 in the heparin group. This finding is consistent with other reports. Observational evidence suggests that the addition of a small amount of heparin overcomes this problem.12 This was permitted during the study, but not applied sufficiently consistently to draw conclusions. A further study to determine the optimum amount of heparin is needed. The question arises of how taurolidine-citrate compares with other antimicrobial lock solutions. Desirable characteristics of taurolidine include effectiveness against bacteria and fungi and the absence of a reported case of microbial resistance10 or side effects after intravenous administration.24,25 For these reasons, we believe taurolidine-citrate would be preferable to an antibiotic lock if the need for thrombolysis can be decreased. Another alternative is high-concentration trisodium citrate (usually 30%), which was reported in a double-blind randomized trial to decrease bacteremia without affecting catheter occlusion.22 However, an open-label trial using 46.7% sodium citrate showed no decrease in bacteremia and an increased incidence of catheter thrombosis,26 and there is a report of life-threatening arrhythmia when 46.7% trisodium citrate was injected rapidly at full strength intravenously. This was the subject of a US Food and Drug Administration warning.27 There may be an increased incidence of paresthesia, metallic taste, and tingling in the fingers associated with its use,26,28,29 although this is not universal.22 The
1067
relative merits of trisodium citrate2 and taurolidine-citrate are uncertain. In conclusion, we observed no difference in the primary outcome of time to first bacteremia or Gram-positive infections; however, our study was not specific to catheter-related bacteremia. It is only realistic to expect taurolidine-citrate to influence infections of intraluminal origin. There was a decrease in bacteremia caused by Gramnegative organisms that suggests that taurolidinecitrate catheter locks may prevent infection with environmental organisms. This warrants further study. The benefit is offset partly by a greater need for thrombolytic therapy. A formulation with better anticoagulant properties is needed.
ACKNOWLEDGEMENTS We acknowledge the support of Mrs Andrea Ashton, Chief Pharmacy Technician for Clinical Trials, Lancashire Teaching Hospitals Trust; the colleagues who have allowed us to investigate patients under their care; and the patients who participated in the study. Support: This work was supported in part by a grant from the Preston branch of the North West Kidney Research Association and a grant from the Liverpool Regional Dialysis Unit Fund. Financial Disclosure: The authors declare that they have no relevant financial interests.
REFERENCES 1. Centers for Disease Control and Prevention. Invasive methicillin-resistant Staphylococcus aureus infections among dialysis patients—United States, 2005. MMWR Morbid Mortal Wkly Rep. 2007;56(9):197-199. 2. McIntyre CW, Hulme LJ, Taal M, Fluck RJ. Locking of tunnelled hemodialysis catheters with gentamicin and heparin. Kidney Int. 2004;66:801-805. 3. Dogra GK, Herson H, Hutchison B, et al. Prevention of tunnelled hemodialysis catheter-related infections using catheter-restricted filling with gentamicin and citrate: a randomised control study. J Am Soc Nephrol. 2002;13:2133-2139. 4. Saxena AK, Panhotra BR, Sundaram DS, Naguib M, Morsy F, Al-Arabi Al-Ghamdi AM. Enhancing the survival of tunnelled hemodialysis catheters using an antibiotic lock in the elderly: a randomised, double-blind clinical trial. Nephrology. 2006;11:299-305. 5. 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(4):596-605. 6. Labriola L, Crott R, Jadoul M. Preventing hemodialysis catheter-related bacteremia with an antimicrobial lock solution: a meta-analysis of prospective randomized trials. Nephrol Dial Transplant. 2008;23:1666-1672. 7. Sweet S, Gobeille S, Landry D, Haessler S, Vaidya C, Braden G. Emergence of gentamicin-resistant bacteremia in hemodialysis patients receiving gentamicin lock
1068 catheter prophylaxis [abstract 206]. Am J Kidney Dis. 2009; 53(4):74A. 8. Dixon J, Steele M, Makanjuola D. Does the use of line locks lead to antibiotic resistant organisms as compared with systemic antibiotics alone in the treatment of hemodialysis catheter infections? Abstract presented at Renal Association Meeting; April 21-24, 2009; Liverpool, UK. http://www. renal.org/pages/media/Meetings/RA%202009%20ABSTRACT_ BOOK.pdf. Accessed December 30, 2009. 9. Torres-Viera C, Thauvin-Eliopoulos C, Souli M, et al. Activities of taurolidine in vitro and in experimental enterococcal endocarditis. Antimicrob Agents Chemother. 2000; 44(6):1720-1724. 10. Shah CB, Mittelam MW, Costerton JW, et al. Antimicrobial activity of a novel catheter lock solution. Antimicrob Agents Chemother. 2003;46(6):1674-1679. 11. Quarello F, Forneris G. Prevention of hemodialysis catheter-related bloodstream infection using an antimicrobial lock. Blood Purif. 2002;20:87-92. 12. Taylor C, Cahill J, Gerrish M, Little J. A new hemodialysis catheter-locking agent reduces infections in hemodialysis patients. J Ren Care. 2008;34(3):116-120. 13. Jurewitsch B, Jeejeebhoy KN. Taurolidine lock: the key to prevention of recurrent catheter-related bloodstream infections. Clin Nutr. 2005;24:462-465. 14. Simon A, Ammann RA, Wiszniewsky G, Bode U, Fleischhack G, Besuden MM. Taurolidine-citrate lock solution (Taurolock) significantly reduces CVAD-associated gram positive infections in pediatric cancer patients. BMC Infect Dis. 2008;8:102-109. 15. Koldehoff M, Zakrzewski JL. Taurolidine is effective in the treatment of central venous catheter-related bloodstream infections in cancer patients. Int J Antimicrob Agents. 2004;24(5):491-495. 16. Allon M. Prophylaxis against dialysis catheterrelated bacteremia with a novel antimicrobial lock solution. Clin Infect Dis. 2003;36:1539-1544. 17. Betjes MGH, van Agteren M. Prevention of dialysis catheter-related sepsis with a citrate-taurolidine-containing lock solution. Nephrol Dial Transplant. 2004;19:1546-1551. 18. Cheesbrough JS, Finch RG, Burden RP. A prospective study of the mechanisms of infection associated with hemodialysis catheters. J Infect Dis. 1986;154(4):579-589.
Solomon et al 19. Johnson DW, MacGinley R, Kay TD et al. A randomised controlled trial of topical exit-site mupirocin application in patients with tunnelled cuffed haemodialysis catheters. Nephrol Dial Transplant. 2002;17:1802-1807. 20. Lok CE, Stanley KE, Hux JE, Richardson R, Tobe SW, Conly J. Hemodialysis infection prevention with polysporin ointment. J Am Soc Nephrol. 2003;13:169-179. 21. Cheesbrough JS, Elliot TJS, Finch RG. A morphological study of bacterial colonisation of intravenous cannulae. J Med Microbiol. 1985;19:149-157. 22. Weijmer MC, Debets-Ossenkopp YJ, Van de Vondervoort FJ, et al. Randomized, clinical trial comparison of trisodium citrate 30% and heparin as catheter-locking solution in hemodialysis patients. J Am Soc Nephrol. 2005;16: 2769-2777. 23. Department of Health. Saving Lives: Reducing Infection, Delivering Clean and Safe Care, High Impact Intervention Nos. 2c and 3, Renal Dialysis Catheter Care Bundle. Crown Publications, UK:Oct;2007. www.clean-safe-care. nhs.uk/toolfiles/19_SL_HII_3_v2.pdf. 24. Willatts SM, Radford S, Leitermann M. Effect of the anti-endotoxic agent, taurolidine, in the treatment of sepsis syndrome, a placebo controlled double-blind controlled trial. Crit Care Med. 1995;23(6):1033-1039. 25. Johnston DA, Phillips G, Perry M, McAlpine H, Richards J, Pennington CR. Taurolin for the prevention of parenteral nutrition related infection: antimicrobial activity and long-term use. Clin Nutr. 1993;12:365-368. 26. Power A, Duncan N, Singh SK, et al. Sodium citrate versus heparin catheter locks for cuffed central venous catheters: a single center randomised control trial of sodium citrate versus heparin line locks for cuffed central venous catheters. Am J Kidney Dis. 2009;53(6):1034-1041. 27. US Food and Drug Administration. FDA Issues Warning on Tricitrasol Dialysis Catheter Anticoagulant. FDA Talk Paper TOO-16 14; April 14, 2000. 28. Ash SR, Mankus RA, Sutton JM. Concentrated sodium citrate (23%) for catheter lock. Haemdial Int. 2000;4: 22-31. 29. Stas KJF, Vanwalleghem J, De Moor B, Keulers H. Trisodium citrate 30% vs heparin 5% as catheter lock in the interdialytic period in twin- or double-lumen dialysis catheters for intermittent hemodialysis. Nephrol Dial Transplant. 2001;16:1521-1522.
I
nfection is a major cause of morbidity and mortality in hemodialysis patients.1 Although decreasing the number of intravascular catheters and increasing the proportion of arteriovenous fistulas decrease the incidence of infection, catheters cannot be eliminated entirely. Pre-
vention of bacteremia is crucial. One approach is to use catheter locks containing an antibiotic. Decreased bacteremia has been reported using gentamicin, cefotaxime, and minocycline and confirmed using meta-analysis.2-6 Risks include toxicity from leakage of the antibiotic into the
From the 1Renal Unit and 2Department of Microbiology, Lancashire Teaching Hospitals, Royal Preston Hospital, Lancashire; 3Renal Unit and 4Health Protection Agency, Manchester Royal Infirmary, Manchester; 5Renal Unit, Royal Liverpool Hospitals, Liverpool; and 6CHICAS, Division of Medicine, School of Health and Medicine, University of Lancaster, Lancaster, UK. Received May 23, 2009. Accepted in revised form November 27, 2009. Originally published online as doi:10.1053/ j.ajkd.2009.11.025 on March 8, 2010.
Trial registration: www.isrctn.org; study number: ISRCTN51902376. Address correspondence to Laurie R. Solomon, FRCP, Renal Unit, Lancashire Teaching Hospitals, Royal Preston Hospital, Sharoe Green Lane, Fulwood, Preston, Lancashire PR2 9HT, UK. Email: [email protected] © 2010 by the National Kidney Foundation, Inc. 0272-6386/10/5506-0013$36.00/0 doi:10.1053/j.ajkd.2009.11.025
1060
American Journal of Kidney Diseases, Vol 55, No 6 (June), 2010: pp 1060-1068
Taurolidine-Citrate Catheter Locks
circulation and the development of antibiotic resistance.3,7,8 An alternative is taurolidine-citrate. Taurolidine (2H-1,2,4-thiadiazine-4,4=-methylenebis[tetrahydro-1,1,1=,1=-tetraoxide]) yields methylol tauranimide components that bind to bacterial and fungal cell walls, causing irreversible damage. This confers very broad-spectrum antimicrobial activity and decreases the development of biofilm.9 Resistance has not been reported and is unlikely because the mode of action resembles a disinfectant, rather than an antibiotic, which exploits a single metabolic pathway.10 Uncontrolled studies suggest that taurolidine locks result in a very low bacteremia rate in hemodialysis and nutrition catheters,11-13 decrease infection in central venous access in children,14 and eradicate catheter-related bloodstream infections unresponsive to systemic antibiotic therapy.15 One nonrandomized controlled trial observed higher bacteremia-free survival at 90 days with taurolidine-citrate than with heparin, but a greater requirement for thrombolytics to maintain catheter patency. This was undertaken in established catheters, which may have contained bacterial biofilm or intraluminal deposits of blood component origin.16 Increased thrombosis might not have occurred if taurolidinecitrate treatment had been started at catheter insertion. Another open-label study randomly assigned patients with temporary and tunneled catheters to taurolidine-citrate or heparin and observed no catheter-related sepsis with taurolidine-citrate compared with 4 episodes with heparin. This study included few tunneled catheters and no information about patency.17 There has been no trial testing taurolidine-citrate from the time of catheter insertion. We undertook a randomized double-blind trial comparing taurolidine-citrate catheter locks with heparin catheter locks started at the time of catheter insertion in hemodialysis patients using tunneled cuffed intravascular catheters. The objective was to determine the incidence of bacteremia and determine whether taurolidine-citrate use is associated with a greater need for thrombolysis.
METHODS Patients The study was undertaken in Northwest England. Catheters were placed at 3 centers and patients were dialyzed at
1061 13 centers. Recruitment took place between November 2006 and May 2008. Adult patients aged ⬎18 years receiving tunneled intravascular catheters for hemodialysis and able to give informed consent were eligible. The Cumbria and Lancashire B Research Ethics committee approved the study (06/Q1309/30), and all patients gave written consent.
Study Procedures Independent pharmacists undertook randomization on a 1:1 basis using computer-generated randomized permuted blocks of 10 patients, which were stratified among the 3 main centers. Physicians responsible for the patient’s care undertook recruitment and obtained consent. Ampoules were labeled “trial catheter lock” and with the patient’s name. All study personnel and participants were blinded to treatment assignment throughout the study until the database was complete. Patients were randomly assigned to receive either taurolidine-citrate or unfractionated heparin (5,000 U/mL) at catheter insertion and after every dialysis treatment until a trial end point. The preferred site of insertion was the right followed by the left internal jugular vein. Right-sided catheters were inserted in an operating room or purpose-built treatment room by a small number of skilled operators. Leftand some right-sided catheters were placed using fluoroscopic guidance in the Radiology Department. Antibiotic prophylaxis was not given. Routine antimicrobial body washes were introduced during the course of the study. Taurolidine-citrate (Taurolock; a solution of 1.35% taurolidine and 4% citrate; Tauropharm AG, www.tauropharm.de) and heparin (porcine heparin sodium [European Pharmacopoeia] at 5,000 U/mL with 1% benzylic acid as preservative [B. Braun AG, www.bbraun.com]) were supplied in identical ampoules. Nurses drew trial solution into a syringe after dialysis and injected the exact locking volume into the catheter lumens. Compliance was checked every 2 weeks. If there was a lapse in the use of trial locks, a note was made and the patient continued in the trial using the trial solution. Exit sites were inspected after every dialysis treatment, cleaned weekly with 2% chlorhexidine in 70% isopropyl alcohol, and covered with a see-through semipermeable membrane (IV 3000 1-HAND; Smith and Nephew, www. smith-nephew.com). If a catheter was replaced, patients could continue in the same arm of the study. Reasons for discontinuation were catheter removal for any reason, kidney transplant, conversion to peritoneal dialysis therapy, transfer to a nonparticipating dialysis unit, development of heparin-induced thrombocytopenia, death, or withdrawal from the study by the patient or attending physician.
Diagnosis and Treatment of Bacteremia The decision to obtain blood cultures was based on symptoms of infection, such as fever (temperature ⬎37.5°C) or rigors associated with dialysis. A single positive blood culture bottle defined a bacteremic episode. Blood culture samples were obtained using aseptic technique from the arterial port of the dialysis circuit. If the same organism was isolated in cultures ⬍3 weeks apart, it was considered the same infection. When different organisms were identified or
1062 if cultures were ⬎3 weeks apart, these were considered separate episodes. Vancomycin and gentamicin (either individually or together) were administered intravenously and as catheter locks for suspected catheter infections according to established protocols for either 2 or 4 weeks, depending on the organism. Oral flucloxacillin or rifampicin was added for Staphylococcus aureus. Doses were adjusted according to blood concentrations. Antibiotic catheter locks were made up in heparin (final concentration, 5,000 U/mL). Catheter removal was considered if the patient remained unwell for ⬎72 hours after starting antibiotic therapy. Exit-site infection was diagnosed on the basis of discharge, erythema, swelling, tenderness, and isolation of a pathogenic organism from a swab.
Thrombolytic Therapy Thrombolytic therapy was given when dialysis was impossible because of poor or no flow. A locking volume of either urokinase (5,000 U/mL; prepared from urokinase powder supplied as 25,000-unit vials; Syner-KINASE; Syner-Med, www.syner-med.com) or tissue plasminogen activator (1 mg/mL; reconstituted from powder supplied in 10-mg vials; Alteplase; Boehringer Ingelheim, www.boehringer-ingelheim. com) was instilled into both catheter lumens for half an hour before retrying. If flow remained poor, a locking dose of thrombolytic was left in the catheter throughout the next interdialytic period. Three doses could be given after 1 episode before considering other options, such as catheter removal. No pumping was used. After successful thrombolysis, attending physicians could add 0.5 mL of 5,000-U/mL heparin to each 5 mL of trial lock solution. This gave a final concentration of 455 U/mL of heparin in the taurolidine-citrate arm, but left the heparin arm unchanged at 5,000 U/mL.
Electron Microscopy of Removed Catheters Two catheter segments (from the tunnel and tip) were cut aseptically and fixed in formalin and osmium tetroxide. After dehydration, catheter sections were immersed in propylene oxide, which caused the catheter material to swell and adherent material to detach. Detached tissue was embedded in resin. Ultrathin sections were cut and stained before examination under a Philips CM10 electron microscope (FEI, www.fei.com).
Solomon et al yes/no), analysis was based on time to first bacteremia event to obtain a more efficient analysis. We report P values and confidence intervals (CIs) for hazard ratios.
Primary and Secondary Outcomes The primary outcome was time to first bacteremia event. Secondary analyses were undertaken for the total number of bacteremic episodes and Gram-positive and Gram-negative infections.
Statistical Analysis A Cox proportional hazards model was used for time to first bacteremia and time to first use of thrombolytic therapy, in each case estimating and testing the treatment effect after stratifying by the 3 main centers and adjusting for age. When calculating catheter survival, recovery of kidney function, kidney transplant, conversion to peritoneal dialysis therapy, availability of alternative vascular access, and transfer to another dialysis unit were considered favorable outcomes. Infection, occlusion, death, and patient or physician choice to leave the study were considered unfavorable. Fisher exact test was used for binary (yes/no) outcomes, and t test, for continuous outcomes, but implemented using Monte Carlo in the case of dialysis vintage because of its highly nonnormal distribution. Counted outcomes were analyzed as rates of occurrence/1,000 patient-days by fitting a Poisson log-linear model with the logarithm of patient-days as an offset. P ⬍ 0.05 is considered significant.
RESULTS Of 256 patients who had tunneled catheters placed during the period of the trial, 110 were recruited (55 taurolidine-citrate, 55 heparin; Fig 1). Most exclusions were logistical given the short time between the decision to insert a catheter and placement or because the patient lacked capacity to consent. At center C, 2 patients were
Sample Size Preliminary data from center A showed that bacteremia occurred in ⬃50% of patients with catheters after 12 months. The all-cause bacteremia rate was 3.3 events/1,000 patientdays for all organisms. We calculated that 31 patients in each group would be needed to show a decrease in the incidence of bacteremia from 50% in the heparin group to 10% in the taurolidine group with a power of 0.9 at the 0.05% level of significance (Fisher exact test). It was believed that allowance should be made for bias of published results and improvements in care during the course of the study. It was decided to include 50 patients in each group and continue 6 months after recruiting the last patient. Although the study was powered on the basis of a binary outcome (bacteremia,
Figure 1. Trial flow diagram (intention to treat analysis). Table 4 lists details of follow-up.
Taurolidine-Citrate Catheter Locks
1063
Table 1. Demographic Details by Type of Catheter Lock TaurolidineCitrate
Heparin
No. of patients
53
54
No. of catheters
56
58
Age (y)
P
0.9
59.8 ⫾ 14.7 56.7 ⫾ 17.4 0.3
Men
26
41
0.009
Race White Asian Black Mixed/other
47 4 1 1
49 5 0 0
0.5a
0 (0-13)
0 (0-6)
0.4
5 7 5 6
3 5 7 11
30
28
Dialysis vintage (y) Cause of ESRD Cystic kidney disease Diabetic nephropathy Glomerulonephritis Hypertensive/ ischemic Other
Note: Values expressed as mean ⫾ standard deviation, median (range), or number. Abbreviation: ESRD, end-stage renal disease. a P is for white versus other.
excluded because the responsible physician considered that they were at high risk of infection because of previous infections and that they lived in an unhygienic environment. Baseline characteristics are listed in Table 1. At the 3 centers, 70, 20, and 20 patients were recruited, representing 62%, 33%, and 33% of catheters placed during the study period, respectively. There was no significant difference between the 2 groups in catheter type (P ⫽ 0.8) or location (P ⫽ 0.9; Table 2). Three patients, 2 in the taurolidine-citrate group and 1 in the heparin group, were allocated trial numbers, but never needed dialysis. Thus, an intention-totreat analysis was performed on 53 patients in the taurolidine-citrate group and 54 in the heparin group. We estimated that trial locks were administered ⬎95% of the time. One patient in the taurolidine-citrate arm had a positive blood culture result for Aerococcus urinate, but was documented before unblinding to have lapsed trial locks for 3 weeks beforehand. Lapse of protocol resulted in default use of heparin locks. This patient was included in the taurolidine-citrate
arm of the intention-to-treat analysis and the heparin arm of the as-treated analysis. Seven patients (3 taurolidine-citrate, 4 heparin) had 2 catheters. Reasons for removal of the first catheter were occlusion (1 heparin, 3 taurolidinecitrate), catheter fell out (1 heparin), infection (1 heparin), and exit-site infection (1 heparin). There were 17,771 catheter-days (8,129 taurolidine-citrate, 9,642 heparin). There were significantly more men in the heparin group despite randomization, but other parameters were similar between the 2 groups (Table 1). Time to First Bacteremia Intention-to-treat analysis showed no significant difference between taurolidine-citrate and heparin with respect to time to first bacteremia after censoring for favorable outcomes (Fig 2). The estimated hazard ratio in favor of taurolidinecitrate was 0.66 (95% CI, 0.28-1.59; P ⫽ 0.4). Age, sex and race were not significant (P ⫽ 0.4, 0.1 and 0.5, respectively). Years on dialysis therapy was borderline significant (P ⫽ 0.05). The secondary as-treated analysis showed no significant difference for taurolidine-citrate over heparin with respect to time to first bacteremia. The estimated hazard ratio in favor of taurolidinecitrate was 0.55 (95% CI, 0.23-1.35; P ⫽ 0.2). Table 2. Catheter Type and Location by Type of Catheter Lock Taurolidine-Citrate
Heparin
53
54
56
58
First catheter type Ash Split Cathb Bard HemoSplitc Circle-Cd Bard HemoGlidec Quinton-Permcathe Bio-Flex Tesiob
31 1 2 5 3 14
34 1 2 5 1 15
Location Internal jugular Subclavian
55 1
57 1
No. of patients No. of catheters a
a
All were cuffed catheters. Medcomp: www.medcompnet.com. c Bard: www.barduk.com. d Horizon Medical Products: www.opitsourcebook.com. e Kendall: www.kendallhq.com. b
1064
Solomon et al
and 6 in the heparin group (P ⫽ 0.9). In the taurolidine-citrate group, all were S aureus. In the heparin group, 4 were S aureus and 2 were coliforms. Four of the 13 infections were associated with positive blood cultures for the same organism (2 heparin, 2 taurolidine-citrate).
Figure 2. Time to first bacteremia. Patients were censored for catheter removal because of recovery of kidney function, kidney transplant, conversion to peritoneal dialysis therapy, availability of alternative vascular access, irreversible occlusion, exit-site infection (without bacteremia), or death or for withdrawal from the trial as a result of patient or physician choice, transfer to a nonparticipating dialysis unit, or development of heparin-induced thrombocytopenia. Abbreviation: TC, taurolidine-citrate.
Bacteremic Episodes There were 11 bacteremic episodes in 9 patients in the taurolidine-citrate group compared with 23 episodes in 16 patients in the heparin group (Table 3). In the heparin group, 3 patients experienced 2 bacteremic episodes and 2 patients experienced 3 bacteremic episodes. In the taurolidine-citrate group, 2 patients experienced 2 bacteremic episodes. The second and third organisms were different species in 6 patients. Blood cultures from 1 patient in the heparin group grew a coagulase-negative staphylococcus with the same antibiogram after a 6-week interval. Nine different species were isolated from patients using heparin compared with 3 different species from patients using taurolidine-citrate. There were 11 and 2 Gram-negative infections in the heparin and taurolidine-citrate groups, respectively. The infection rate for Gram-negative organisms was significantly lower in the taurolidinecitrate group compared with the heparin group (P ⫽ 0.02), but there was no difference in the number of Gram-positive infections (P ⫽ 0.8; Table 3). Exit Sites Pathogenic organisms were cultured from exit sites on 13 occasions. There were 7 positive culture results in the taurolidine-citrate group
Use of Thrombolytic Therapy More patients in the taurolidine-citrate group needed thromobolytic therapy at least once (taurolidine-citrate, 28 of 53; heparin, 14 of 54; P ⫽ 0.006). There was a significant adverse effect of taurolidine-citrate with respect to time to first use of Table 3. Number of Bacteremic Episodes and Organisms Isolated
No. of patients a
Bacteremic episodes Gram positivea,b Staphylococcus aureus (MSSA) Staphylococcus aureus (MRSA) Other Gram positive Gram negativea,d No. of infections/patient 1 2 3 No. of catheter-days
TaurolidineCitrate
Heparin
53
54
11 (1.4) 9 (1.1) 5
23 (2.4) 12 (1.2) 6
1c
2
3 2 (0.2)
P
0.1 0.8
4 11c (1.1) 0.02 0.3
7 2 0
11 3 2
8,129
9,642
Note: Based on intention-to-treat analysis. Abbreviations: MRSA, methicillin-resistant S aureus; MSSA, methicillin-sensitive S aureus. a Figures outside brackets are actual numbers. Figures in brackets are per 1,000 catheter-days. b Staphylococcus aureus (taurolidine citrate, 6; heparin, 8), coagulase-negative Staphylococcus (taurolidine citrate, 2; heparin, 4), Aerococcus urinae (taurolidine citrate, 1). c On 2 occasions, 2 organisms were grown from the same blood culture. One patient in the heparin group had Klebsiella and Enterobacter asbiriae; 1 patient in the taurolidine-citrate group had methicillin-resistant S aureus and coagulase-negative staphylococcus from the same blood culture. The second organism is not included in this table. d Escherichia coli (taurolidine citrate, 2; heparin, 2), Klebsiella oxytoca (heparin, 1), Delftia acidovorans (heparin, 1), Enterobacter aerogenes (heparin, 1), Haemophilus parainfluenzae (heparin, 1), Serratia marcescens or S liquefaciens (heparin, 4), and Strenophonomonas maltophilia (heparin, 1).
Taurolidine-Citrate Catheter Locks
1065
infection (P ⫽ 0.8). Eight taurolidine-citrate and 3 heparin catheters were removed because of catheter occlusion (P ⫽ 0.06). There were 11 deaths in the taurolidine-citrate group compared with 8 deaths in the heparin group (P ⫽ 0.2). Differences Between Centers
Figure 3. Time to first use of thrombolytic therapy. Patients were censored for catheter removal because of recovery of kidney function, kidney transplant, conversion to peritoneal dialysis therapy, availability of alternative vascular access, bacteremia, exit-site infection, or death or for withdrawal from the trial as a result of patient or physician choice, transfer to a nonparticipating dialysis unit, or development of heparin induced thrombocytopenia. Abbreviation: TC, taurolidine-citrate.
thrombolytic therapy (Fig 3). The estimated hazard ratio against taurolidine-citrate was 2.5 (95% CI, 1.3-5.2; P ⫽ 0.008). In addition, 8 catheters in the taurolidinecitrate group and 3 in the heparin group were lost because of occlusion (P ⫽ 0.06). There were 79 administrations of thrombolytic to the taurolidinecitrate group and 38 to the heparin group. Thus, thrombolysis was successful ⬎90% of the time. Catheter Survival Median survival for first catheters was 271 days (95% CI, 245-297) for taurolidine-citrate and 358 days (95% CI, 270-445) for heparin after censoring for favorable outcomes, but including all deaths and withdrawals for patient or physician choice as adverse outcomes (P ⫽ 0.3). Reasons for Trial Termination Thirty-two taurolidine-citrate and 36 heparin catheters were removed for reasons unrelated to the trial, including alternative access, peritoneal dialysis therapy, recovery of kidney function, kidney transplant, transfer to another dialysis unit, and physician or patient choice (Table 4). Two taurolidine-citrate and 8 heparin catheters were removed because of persistent signs of infection after a positive blood culture (P ⫽ 0.1). Two taurolidine-citrate and 3 heparin catheters were removed because of exit-site or tunnel
Subgroup analysis was undertaken for center A, which had provided the background data for sample-size calculation, contributed most study participants, and recruited the greatest proportion of eligible patients, patient-days, and bacteremic episodes. Analysis was undertaken because of possible selection of lower risk patients from centers B and C, which had low rates of infection in the heparin arm (0 and 1.51/1,000 catheter-days). Center A contributed 7 bacteremic episodes in the taurolidine-citrate arm and 21 in the heparin arm (P ⫽ 0.02). Estimated hazard ratios for time to first bacteremia were 0.5 (95% CI, 0.2-1.2; P ⫽ 0.12) in the intention-totreat analysis and 0.37 (95% CI, 0.13-1.02; P ⫽ Table 4. Reasons for Catheter Removal or Trial Termination
No. of patients No. of catheters Bacteremia Exit-site infection Removal for occlusion/flow problems Patient choice Physician choice Heparin-induced thrombocytopenia Recovered renal function Alternative access available Conversion to peritoneal dialysis Transplant Transfer to another dialysis unit Reached end of trial Catheter fell out Deaths
TaurolidineCitrate
Heparin
53 56 2 2 8
54 58 8 3 3
0.1 0.8 0.1
2 1 1
2 0 0
0.8 0.2 0.2
4 17 3
4 15 1
0.7 0.3 0.2
1 3
2 4
0.7 0.9
1 0 11a
7 1 8b
0.05 0.3 0.2
P
a Cause of death unknown (6), cardiac (2), staphylococcus sepsis (1), cirrhosis (1), miliary tuberculosis (1), and myeloma (1). b Cause of death unknown (4), infected toe (1), mesothelioma (1), myeloma (1), and peripheral ischemia (1).
1066
0.06) in the as-treated analysis in favor of taurolidine-citrate. Intraluminal Material at Time of Removal Fourteen catheters (4 taurolidine-citrate, 10 heparin) were examined using transmission electron microscopy. All showed deposition of material derived from red blood cells, fibrin, and platelets at the tip. In the tunnel segment where the locking solution may be maintained without blood dilution between hemodialysis sessions, all except 4 catheters showed similar deposits. There were bacteria in 2 catheters: 1 in the heparin arm had an S aureus bacteremia, and organisms resembling Gram-positive cocci surrounded by polymorphs were seen in the tunnel and tip segments; in 1 catheter in the taurolidinecitrate arm, 1 Gram-negative rod was seen among red blood cells at the tip, but this was not associated with clinical or microbiological evidence of infection. There was no bacterial biofilm in any catheter.
DISCUSSION In this randomized double-blind controlled trial of taurolidine-citrate catheter locks in hemodialysis patients, we found no difference in time to first infection, but a significant decrease in the number of Gram-negative infections. There was a greater need for thrombolytic therapy in taurolidine-citrate locks compared with heparin locks. There was no significant difference in catheter survival. Results are consistent with previous open-label trials and observational studies.12,16,17 The distribution of Gram-negative organisms was skewed: both infections in the taurolidinecitrate arm were with typical endogenous flora, whereas 9 in the heparin arm were caused by organisms often associated with environmental reservoirs. This result suggests that there was protection against bacterial contamination during connect-disconnect procedures. However, there was no difference in the incidence of Grampositive organisms, including S aureus. In vitro studies indicate that the concentration of taurolidine was sufficient to kill S aureus.9,10 One possibility is that S aureus ingresses by a different portal, such as the exit site. We decided to include all positive blood cultures without attempting to differentiate catheter-related blood-
Solomon et al
stream infections from those originating at other foci because proving that bacteremia is catheter related requires simultaneous blood cultures from a peripheral vein and the catheter. We wanted to avoid unnecessary venipuncture and continue the standard practices of the participating units. A positive blood culture was a clear and verifiable measure. Some patients in both arms were very ill for other reasons. One of 2 Gramnegative infections in the taurolidine-citrate group occurred in a patient with cirrhosis and spontaneous bacterial peritonitis. Four episodes of bacteremia, 2 in each group, were associated with exit-site infection. The finding of approximately equal numbers of positive swabs from exit sites in the 2 groups is similar to a previous report17 and suggests that exit-site infections have a different pathogenesis. Bacterial infections related to dialysis catheters can develop through extra- and intraluminal routes.18 Catheter locks combat intraluminal contamination and cannot be expected to eliminate bacteremia originating in the exit site. This may require different measures. Mupirocin or polysporin cream applied to the exit site may be effective prophylaxis.19,20 We failed to show bacterial glycocalyx in any catheter. We found accumulation of fibrin, erythrocytes, and platelets at the tip of all catheters in both groups. Therefore, it was not possible to determine whether taurolidine-citrate destroys established bacterial biofilm. Bacterial biofilm may not be universal.21 This warrants further study. The relative decrease in bacteremia should be viewed in light of the incidence in the heparin arm because a low rate of infection in the control group reduces the scope for improvement. The relative decrease using taurolidine-citrate in our study was less than the 90% decrease in bacteremia reported by Allon,16 but the rate in the heparin arm in that study was 5.6 events/1,000 catheter-days (compared with 2.4 events/1,000 catheter-days in our trial). A higher incidence also was observed in the heparin arms of studies using gentamicin and 30% citrate.2,22 During the study period, the UK Department of Health encouraged rigorous infection control discipline to decrease methicillin-resistant S aureus infection.23 This may have contributed to a low incidence of bacteremia in our study.
Taurolidine-Citrate Catheter Locks
We performed subgroup analysis at center A, which suggested a significant decrease in the rate of infection and trend to decreased hazard ratio for time to first infection. Center A contributed most patients and catheters placed during the study period. It is possible that this was more representative of the general dialysis population and that the other 2 centers recruited lower risk patients. Furthermore, fewer taurolidine-citrate patients had ⬎1 bacteremic episode. Taurolidinecitrate therefore may be of greatest benefit in patients at high risk of infection. The main disadvantage of taurolidine-citrate was a greater need for thrombolytic therapy to maintain catheter patency. This was successful in most cases; however, catheter removal was required in 8 patients in the taurolidine-citrate group compared with 3 in the heparin group. This finding is consistent with other reports. Observational evidence suggests that the addition of a small amount of heparin overcomes this problem.12 This was permitted during the study, but not applied sufficiently consistently to draw conclusions. A further study to determine the optimum amount of heparin is needed. The question arises of how taurolidine-citrate compares with other antimicrobial lock solutions. Desirable characteristics of taurolidine include effectiveness against bacteria and fungi and the absence of a reported case of microbial resistance10 or side effects after intravenous administration.24,25 For these reasons, we believe taurolidine-citrate would be preferable to an antibiotic lock if the need for thrombolysis can be decreased. Another alternative is high-concentration trisodium citrate (usually 30%), which was reported in a double-blind randomized trial to decrease bacteremia without affecting catheter occlusion.22 However, an open-label trial using 46.7% sodium citrate showed no decrease in bacteremia and an increased incidence of catheter thrombosis,26 and there is a report of life-threatening arrhythmia when 46.7% trisodium citrate was injected rapidly at full strength intravenously. This was the subject of a US Food and Drug Administration warning.27 There may be an increased incidence of paresthesia, metallic taste, and tingling in the fingers associated with its use,26,28,29 although this is not universal.22 The
1067
relative merits of trisodium citrate2 and taurolidine-citrate are uncertain. In conclusion, we observed no difference in the primary outcome of time to first bacteremia or Gram-positive infections; however, our study was not specific to catheter-related bacteremia. It is only realistic to expect taurolidine-citrate to influence infections of intraluminal origin. There was a decrease in bacteremia caused by Gramnegative organisms that suggests that taurolidinecitrate catheter locks may prevent infection with environmental organisms. This warrants further study. The benefit is offset partly by a greater need for thrombolytic therapy. A formulation with better anticoagulant properties is needed.
ACKNOWLEDGEMENTS We acknowledge the support of Mrs Andrea Ashton, Chief Pharmacy Technician for Clinical Trials, Lancashire Teaching Hospitals Trust; the colleagues who have allowed us to investigate patients under their care; and the patients who participated in the study. Support: This work was supported in part by a grant from the Preston branch of the North West Kidney Research Association and a grant from the Liverpool Regional Dialysis Unit Fund. Financial Disclosure: The authors declare that they have no relevant financial interests.
REFERENCES 1. Centers for Disease Control and Prevention. Invasive methicillin-resistant Staphylococcus aureus infections among dialysis patients—United States, 2005. MMWR Morbid Mortal Wkly Rep. 2007;56(9):197-199. 2. McIntyre CW, Hulme LJ, Taal M, Fluck RJ. Locking of tunnelled hemodialysis catheters with gentamicin and heparin. Kidney Int. 2004;66:801-805. 3. Dogra GK, Herson H, Hutchison B, et al. Prevention of tunnelled hemodialysis catheter-related infections using catheter-restricted filling with gentamicin and citrate: a randomised control study. J Am Soc Nephrol. 2002;13:2133-2139. 4. Saxena AK, Panhotra BR, Sundaram DS, Naguib M, Morsy F, Al-Arabi Al-Ghamdi AM. Enhancing the survival of tunnelled hemodialysis catheters using an antibiotic lock in the elderly: a randomised, double-blind clinical trial. Nephrology. 2006;11:299-305. 5. 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(4):596-605. 6. Labriola L, Crott R, Jadoul M. Preventing hemodialysis catheter-related bacteremia with an antimicrobial lock solution: a meta-analysis of prospective randomized trials. Nephrol Dial Transplant. 2008;23:1666-1672. 7. Sweet S, Gobeille S, Landry D, Haessler S, Vaidya C, Braden G. Emergence of gentamicin-resistant bacteremia in hemodialysis patients receiving gentamicin lock
1068 catheter prophylaxis [abstract 206]. Am J Kidney Dis. 2009; 53(4):74A. 8. Dixon J, Steele M, Makanjuola D. Does the use of line locks lead to antibiotic resistant organisms as compared with systemic antibiotics alone in the treatment of hemodialysis catheter infections? Abstract presented at Renal Association Meeting; April 21-24, 2009; Liverpool, UK. http://www. renal.org/pages/media/Meetings/RA%202009%20ABSTRACT_ BOOK.pdf. Accessed December 30, 2009. 9. Torres-Viera C, Thauvin-Eliopoulos C, Souli M, et al. Activities of taurolidine in vitro and in experimental enterococcal endocarditis. Antimicrob Agents Chemother. 2000; 44(6):1720-1724. 10. Shah CB, Mittelam MW, Costerton JW, et al. Antimicrobial activity of a novel catheter lock solution. Antimicrob Agents Chemother. 2003;46(6):1674-1679. 11. Quarello F, Forneris G. Prevention of hemodialysis catheter-related bloodstream infection using an antimicrobial lock. Blood Purif. 2002;20:87-92. 12. Taylor C, Cahill J, Gerrish M, Little J. A new hemodialysis catheter-locking agent reduces infections in hemodialysis patients. J Ren Care. 2008;34(3):116-120. 13. Jurewitsch B, Jeejeebhoy KN. Taurolidine lock: the key to prevention of recurrent catheter-related bloodstream infections. Clin Nutr. 2005;24:462-465. 14. Simon A, Ammann RA, Wiszniewsky G, Bode U, Fleischhack G, Besuden MM. Taurolidine-citrate lock solution (Taurolock) significantly reduces CVAD-associated gram positive infections in pediatric cancer patients. BMC Infect Dis. 2008;8:102-109. 15. Koldehoff M, Zakrzewski JL. Taurolidine is effective in the treatment of central venous catheter-related bloodstream infections in cancer patients. Int J Antimicrob Agents. 2004;24(5):491-495. 16. Allon M. Prophylaxis against dialysis catheterrelated bacteremia with a novel antimicrobial lock solution. Clin Infect Dis. 2003;36:1539-1544. 17. Betjes MGH, van Agteren M. Prevention of dialysis catheter-related sepsis with a citrate-taurolidine-containing lock solution. Nephrol Dial Transplant. 2004;19:1546-1551. 18. Cheesbrough JS, Finch RG, Burden RP. A prospective study of the mechanisms of infection associated with hemodialysis catheters. J Infect Dis. 1986;154(4):579-589.
Solomon et al 19. Johnson DW, MacGinley R, Kay TD et al. A randomised controlled trial of topical exit-site mupirocin application in patients with tunnelled cuffed haemodialysis catheters. Nephrol Dial Transplant. 2002;17:1802-1807. 20. Lok CE, Stanley KE, Hux JE, Richardson R, Tobe SW, Conly J. Hemodialysis infection prevention with polysporin ointment. J Am Soc Nephrol. 2003;13:169-179. 21. Cheesbrough JS, Elliot TJS, Finch RG. A morphological study of bacterial colonisation of intravenous cannulae. J Med Microbiol. 1985;19:149-157. 22. Weijmer MC, Debets-Ossenkopp YJ, Van de Vondervoort FJ, et al. Randomized, clinical trial comparison of trisodium citrate 30% and heparin as catheter-locking solution in hemodialysis patients. J Am Soc Nephrol. 2005;16: 2769-2777. 23. Department of Health. Saving Lives: Reducing Infection, Delivering Clean and Safe Care, High Impact Intervention Nos. 2c and 3, Renal Dialysis Catheter Care Bundle. Crown Publications, UK:Oct;2007. www.clean-safe-care. nhs.uk/toolfiles/19_SL_HII_3_v2.pdf. 24. Willatts SM, Radford S, Leitermann M. Effect of the anti-endotoxic agent, taurolidine, in the treatment of sepsis syndrome, a placebo controlled double-blind controlled trial. Crit Care Med. 1995;23(6):1033-1039. 25. Johnston DA, Phillips G, Perry M, McAlpine H, Richards J, Pennington CR. Taurolin for the prevention of parenteral nutrition related infection: antimicrobial activity and long-term use. Clin Nutr. 1993;12:365-368. 26. Power A, Duncan N, Singh SK, et al. Sodium citrate versus heparin catheter locks for cuffed central venous catheters: a single center randomised control trial of sodium citrate versus heparin line locks for cuffed central venous catheters. Am J Kidney Dis. 2009;53(6):1034-1041. 27. US Food and Drug Administration. FDA Issues Warning on Tricitrasol Dialysis Catheter Anticoagulant. FDA Talk Paper TOO-16 14; April 14, 2000. 28. Ash SR, Mankus RA, Sutton JM. Concentrated sodium citrate (23%) for catheter lock. Haemdial Int. 2000;4: 22-31. 29. Stas KJF, Vanwalleghem J, De Moor B, Keulers H. Trisodium citrate 30% vs heparin 5% as catheter lock in the interdialytic period in twin- or double-lumen dialysis catheters for intermittent hemodialysis. Nephrol Dial Transplant. 2001;16:1521-1522.