Abstract
♦ Background
♦ Objective
♦ Methods
♦ Results
♦ Conclusion
Ion beam-based technologies have led to processes by which surface properties of catheters can be modified without adversely affecting the mechanical properties of the polymer bulk (1). One such process permits implantation of silver ions into the surface of silicone rubber catheters (Spi-Argent II; Spire Corporation, Bedford, Massachusetts, USA). Silver is a known bacteriostatic/bactericidal agent and has plenty of precedents for medical applications. In various in vitro and in vivo studies, silver-treated materials have been effective against Staphylococcus epidermidis, Staphylococcus aureus, Escherichia coli, Enterococcus faecalis, Candida albicans, and Pseudomonas aeruginosa (1-3). In addition, ion implantation reportedly enhances catheter surface characteristics to reduce bacterial attachment and biofilm formation (4). Extensive toxicological tests demonstrate that silver-ion implanted materials are biocompatible (1).
To date, there are no clinical studies evaluating the efficacy of silver-ion implanted peritoneal access devices. The potential benefit of silver-ion treated peritoneal dialysis catheters was extrapolated from favorable experience with intravascular appliances. The aim of the present study was to test the effectiveness of silver-treated catheters in reducing dialysis-related infections in chronic renal failure patients.
Patients and Methods
The study design was a prospective, randomized, controlled clinical trial evaluating the efficacy of silver-ion implanted peritoneal dialysis catheters in reducing dialysis-related infections. The Institutional Review Board approved the study and informed consent was obtained from all patients. Between July 1997 and September 2000, consecutive renal failure patients entering the peritoneal dialysis program at the Kaiser Permanente Bellflower Medical Center were approached to participate in the study. Randomization was performed by random number table allocation to either the silver-ion treated study group or the untreated control group. In June 1999, the manufacturer of the study and control catheters (Sil-Med Corporation, Taunton, Massachusetts, USA) advised the principal investigators that the company was acquired by Saint-Gobain Performance Plastics. As a result of the new company's policy, the peritoneal dialysis catheter product line, along with any products used for more than 30 days in patient care, was discontinued effective immediately. There were no alternative manufacturers of the study catheter. Due to a limited supply of remaining catheters, and these having expiration dates, the study protocol was modified to include patients that refused randomization but would otherwise consent to the provisions of the study protocol. In addition, there were six more control catheters than available study catheters and the rest of the prospective study subjects were consented for participation in the protocol without the provision for randomization.
Both study and control peritoneal catheters were of the Tenckhoff style with a coiled intraperitoneal tip, double cuffs, and a straight intramural segment. All catheters were implanted in a paramedian location using a laparoscopic technique developed by the authors (5). A prophylactic antibiotic, cefazolin, or vancomycin in the event of cephalosporin allergy, was administered prior to all procedures. Daily oral ciprofloxacin was continued for 10 days postoperatively. The deep cuff of the catheter was implanted in the rectus sheath and the superficial cuff within 2 –3 cm of the exit wound. The exit site was directed laterally and slightly below the transverse plane. A chlorhexidine-impregnated wafer (BioPatch; Johnson & Johnson, Arlington, Texas, USA) was left in place around the catheter tubing at the exit wound for 2 weeks, at which time the patient began a routine of daily exit-site cleansing with antibacterial soap and hydrogen peroxide. Patients were permitted to resume showering after 1 month if wound healing was uncomplicated. The exit wound was kept covered with sterile gauze. All implants were allowed at least 2 weeks to heal before instituting dialysis.
A surveillance and treatment program for S. aureus nasal carriage was in place in the peritoneal dialysis unit during the study period (6). Nasal cultures were performed at the time of entry into the dialysis program and quarterly thereafter. Patients with S. aureus carriage underwent treatment with intranasal mupirocin applied twice daily for 5 consecutive days for each of 3 consecutive months.
Patients were generally evaluated on a monthly basis in the clinic and immediately in the event of an acute problem. Exit-site infections were diagnosed if signs of redness and purulent discharge were present (7). Tunnel infection included induration or redness over the subcutaneous course of the catheter associated with tenderness and pain, with or without abscess formation (7). Exit-site and tunnel infections frequently overlapped in clinical presentation and were combined for statistical purposes. Peritonitis was defined clinically as abdominal pain and a cloudy dialysate yielding a leukocyte count greater than 100/mm3 with greater than 50% polymorphonuclear cells (8).
In general, removal of peritoneal catheters for exit-site and tunnel-tract infections was performed if there was a failure to respond within 2–4 weeks to a treatment program that included appropriate antibiotic therapy, intensified exit-wound care, and unroofing of the infected tunnel tract with shaving of the superficial cuff (7). Removal of peritoneal catheters for refractory peritonitis was performed if no improvement was noted within 5 days of starting appropriate antibiotic treatment. Recrudescence of peritonitis within 2 weeks of discontinuing antibiotic therapy led to the removal of peritoneal catheters for relapsing disease (8). Episodes of exit-site infection and peritonitis were counted as separate events if the episode occurred more than 4 weeks after stopping antibiotic therapy or if the infection was caused by a different organism.
The study was designed to detect a 50% reduction in the historical rate of exit-site infections at a one-tailed 0.025 level of significance with 80% power. Our historical rate was 0.433 exit-site infections per patient-year of dialysis. To detect such a reduction required at least 64.7 patient-years per group.
Fisher's exact test was used to compare nominal data. Age and duration of subject observation for the study and control groups were compared using the Wilcoxon rank-sum test. Infection rates were compared by Poisson regression analysis. Antibiotic-free intervals representing the number of days between infections during which a patient was not taking antibiotics for exit-site infection, peritonitis, or both infections combined were compared by the Wilcoxon rank-sum test. Probability distributions for remaining free of infection and overall catheter survival were estimated using the method of Kaplan and Meier. Comparison of probability curves was performed with the log-rank test. All results were considered significant at p less than 0.05.
Results
Between July 1997 and September 2000, 140 consecutive renal failure patients entering the peritoneal dialysis program at the Kaiser Permanente Bellflower Medical Center were approached to participate in the study. One hundred twenty-nine patients were randomized to the study (n = 66) and control groups (n = 63). Five patients entered the study twice, randomized on both occasions. Three of these patients returned to the same group as control subjects and two entered the opposite group. Another 5 patients refused randomization but otherwise consented to study provisions. Of these 5 patients, 3 selected the control catheter (1 patient reportedly had a silver allergy) and 2 selected the study catheter. After exhaustion of the supply of study catheters, 6 patients consented to the study protocol with a control catheter. One randomized patient implanted with a study catheter moved out of the service area before starting dialysis and was excluded. This report includes the second-catheter experience of the 5 patients that had 2 implants and the 11 nonrandomized subjects, since none of the study analyses with or without these participants significantly changed the silver catheter results. The final subject groups were comprised of 67 silver-treated study catheters and 72 control catheters. Data collection continued through September 2001.
Demographic data for the study and control group subjects are shown in Table 1. There was no significant difference between the two groups with respect to age, gender, diabetic status, S. aureus nasal carriage, previous dialysis experience, or duration of observation. The causes of renal failure for the participants in the two groups are shown in Table 2. There was no significant difference in the etiology of end-stage renal disease between the study and control group subjects.
Demographic Data
Based on the Wilcoxon rank-sum test.
Based on Fisher's exact test.
Etiology of End-Stage Renal Disease in the Study and Control Group Subjects
Culture results of exit-site infection episodes for the study and control groups are shown in Table 3. The study group experienced 65 separate exit-site infections in 33 patients; 50 infections occurred in 26 subjects of the control group. Staphylococcus aureus, P. aeruginosa, and coliforms were the commonest isolates from exit-site infections.
Culture Results of Exit-Site Infection Episodes for Study and Control Groups
Culture results of peritonitis episodes in the study and control groups are shown in Table 4. There were 47 separate episodes of peritonitis in 32 patients in the study group. The control group experienced 40 episodes of peritonitis in 27 subjects. Coagulase-negative staphylococcus, S. aureus, and coliforms were the commonest organisms causing peritonitis.
Culture Results of Peritonitis Episodes for Study and Control Groups
Based upon subject numbers and duration of observation, the study as performed was capable of detecting a 40% reduction in exit-site infections and a 30% reduction in peritonitis rate at a one-tailed nominal 0.025 significance level with at least 80% power.
Exit-site infection rates for the study and control groups were 0.52 and 0.45 episodes per patient-year of dialysis respectively. Poisson regression analysis demonstrated no significant difference in the rates (p > 0.4). Peritonitis rates for the study and control groups were identical at 0.37 episodes per patient-year of dialysis. Poisson analysis comparing peritonitis rates of the study and control groups confirmed lack of significant difference (p > 0.9).
Antibiotic-free intervals between infections for the study and control groups were not significantly different for exit-site infections (p = 0.58), peritonitis (p = 0.44), or both infections combined (p = 0.47).
Actuarial analysis showed no significant difference in the probability of remaining free of exit-site infection (Figure 1) or peritonitis (Figure 2) between study and control subjects. Comparison of catheter survival curves for study and control groups showed no significant difference (Figure 3).
Probability curves for remaining free of exit-site infection were not significantly different for the study and control groups (log-rank test: χ2 = 1.598, df = 1, p = 0.21).
Probability curves for remaining free of peritonitis were not significantly different for the study and control groups (log-rank test: χ2 = 0.115, df = 1, p = 0.74).
Catheter survival for the study and control groups was not significantly different (log-rank test: χ2 = 0.282, df = 1, p = 0.6).
Infectious complications that led to catheter removal were due to peritonitis. There were 9 study group catheters and 10 control group catheters removed because of peritonitis. Two subjects in the study group had concurrent exit-site and tunnel-tract infections involving both catheter cuffs. The causes of these two episodes of catheter infection-related peritonitis were S. aureus in one case and P. aeruginosa in the second. The micro-organisms that produced peritonitis and catheter loss in the two groups are itemized in Table 5.
Micro-Organisms That Caused Peritoneal Dialysis Catheter Loss in the Study and Control Group Subjects
Discussion
Surface modifications of implanted medical devices that enhance resistance to infectious complications continue to be of ongoing interest. The known biocidal characteristics of silver have made it a popular investigatory candidate for surface treatment processes. Any assessment of silver treatment of a medical device must examine the methodology of silver attachment and chemical bioavailability. The antibacterial activity of silver depends on the continuous release of silver ions at the catheter surface.
The peritoneal dialysis catheters in the present study were treated with a silver ion beam implantation process designated by the proprietary name SPI-Argent II. Silver ions in the energy range of 50 –100 keV and dose range of 4 - 6 × 1015 ions/cm2 were used to bombard the surface of the catheter. At this energy range, the silver ions penetrated the surface of the catheter to a depth of 200 – 300 nm. Unlike a coating process, no material is added to the surface of the catheter with ion implantation. The reported advantage of ion implantation is that it modifies only the surface of the catheter without affecting the bulk properties of the silicone rubber. In addition to the antimicrobial effects of silver, the interaction of silver ions with the silicone polymer at the catheter surface reportedly produces physical and chemical structural changes that reduce bacterial attachment and biofilm formation (1).
The failure of the silver-ion implanted catheters in the present study to show a clinical effect in reducing infectious complications may be due to the low surface availability of silver. Furthermore, current ion beam technology does not permit treatment of the internal surface of the catheter. It comes as no surprise that the observed peritonitis rates were equal for the study and control groups, since the lumina of catheters are just as susceptible to bacterial biofilm formation as are the external surfaces (9).
It has been suggested that performing silver ion implantation at lower energy levels may improve the ability of the surface treatment to prevent infection (4). Compared to higher energy levels, reducing the implanting energy to 10 keV appears to increase the availability of silver. Penetration depth of silver ions at 10 keV has been calculated at approximately 18.7 nm. In addition, the lower energy level produces structural changes on the polymer surface that are thought to be more effective in reducing bacterial adhesion. The use of higher ion doses, 2 × 1016 ions/cm2, in combination with low implantation energy results in the formation of colloidal silver particles at the catheter surface. This increased accessibility of colloidal silver at the surface may further enhance the antibacterial effect. Further study is required.
Previous animal studies have investigated the use of catheter surface treatment with ion beam-assisted deposition (IBAD) of silver, a process previously known as SPI-Argent I and now referred to simply as SPI-Argent (Spire Corporation). The IBAD process utilizes evaporated silver that is allowed to condense as a thin film on the catheter surface (1). Adhesion of the film to the catheter is facilitated by ion beam bombardment. A silver coating of approximately 0.5 urn thick is added to the surface of the treated material. Just as in the case of ion implantation, the IBAD process treats only the outside surface of the catheter tubing.
Animal studies utilizing silver-IBAD treated peritoneal dialysis catheters tended to support improved wound healing and lower incidences of infection compared to untreated devices (2,10-12). These animal studies were short-term observations, involved small subject numbers, and the results are best characterized as preliminary. There are no published human studies involving the efficacy of IBAD-treated peritoneal dialysis catheters. Vascular catheters treated with the silver IBAD process have demonstrated lower incidences of positive blood and catheter tip cultures in a randomized and controlled study (13). It remains to be seen if results of short-term vascular devices can be extrapolated to long-term peritoneal dialysis catheters. Since ion implantation and IBAD processes are limited to the external surfaces of catheter devices, the best that can be expected from these technologies is a reduction in exit-site infection, tunnel-tract infection, and catheter infection-related peritonitis.
The addition of metallic silver components to peritoneal dialysis access devices has been tried as a means of preventing dialysis-related infections. A short-term animal study demonstrated that a sterling silver disk attached to the anterior abdominal wall protected against catheter infection-related peritonitis up to 4 weeks postimplant (14). However, a prospective, randomized, controlled clinical trial employing a silver ring device mounted on the dialysis catheter at the skin exit site failed to demonstrate clinical efficacy in preventing exit-site infections, tunnel-tract infections, peritonitis, or infection-related catheter losses (15). Device design may have adversely influenced the study results. Displacement of the ring into the tunnel tract contributed to the occurrence of infections. In addition, some patients withdrew from the study because of discomfort with the device.
To address the problem of achieving antimicrobial activity on both the external and luminal surfaces, another approach has been to disperse either silver chloride or metallic silver throughout the polymer matrix (3). Silver was incorporated into the catheter material at a concentration of 0.1% – 0.5% (weight/ volume). Concentrations of silver in the matrix higher than 1% wt/vol have a negative influence on the physical properties of the plastic material. The distribution of submicron silver particles evenly throughout the catheter matrix is thought to create a larger antimicrobially active surface than can be achieved by surface coating or a sheet of metallic silver (16). While in vitro studies have demonstrated that catheters containing silver in microdispersion on the external and internal surfaces have significant antimicrobial activity, clinical tests with silver-treated vascular catheters have not consistently produced favorable results (17,18). Another concern relates to the process of using solvents to swell catheter materials to disperse silver within the polymer bulk. Organic solvents may induce stress cracking within the polymer matrix that can lead to device failure over the long term (19, 20).
New surface-treatment technologies for peritoneal dialysis catheters that can reduce the rate of infectious complications without adversely affecting basic design and function would constitute a major advance in the field. The future goals of biomaterials development are to define processes that will permit long-term release of an optimal concentration of silver ions capable of antimicrobial activity, yet remain nontoxic to the host. The reservoir of silver in the catheter should be sufficient for long-term activity but should not alter the mechanical properties of the catheter material. And finally, the antimicrobial activity must be present on both the external and the internal surfaces of catheter devices.