- Email: [email protected]
A prospective study of hemodialysis access-related bacterial infections
J Infect Chemother (2002) 8:242–246 DOI 10.1007/s10156-002-0184-8
© Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases 2002
ORIGINAL ARTICLE Ibrahiem Saeed Abdulrahman · Samir H. Al-Mueilo Huda A. Bokhary · Gani O.A. Ladipo Abdulla Al-Rubaish
A prospective study of hemodialysis access-related bacterial infections
Received: November 8, 2001 / Accepted: April 25, 2002
Abstract The objective of this study was to describe hemodialysis vascular-access related infections that occurred in hemodialysis patients over an 18-month period. The study is a prospective descriptive analysis of incidence infection rates in a hemodialysis unit in a tertiary-care medical center. Prospective surveillance for hemodialysis vascular accessrelated infection was performed for all patients undergoing hemodialysis from November 1999 through April 2001 at King Fahd Hospital of King Faisal University, Al-Khobar, Saudi Arabia. The total number of dialysis sessions was calculated. The type of vascular access was noted. Cultures were obtained and all infections were recorded and infection rates were calculated. There were 9627 hemodialysis sessions (5437 via permanent fistulae or grafts, 2409 via temporary central catheters, and 1781 via permanent tunneled catheters) during the 18-month study period. We identified a total of 109 infections, for a rate of 11.32/1000 dialysis sessions (ds). Of the 109, 23 involved permanent fistulae or grafts (4.23/1000 ds); 18 involved permanenttunneled central catheter infections (10.1/1000 ds); and 68 involved temporary-catheter infections (28.23/1000 ds). There were 38 bloodstream infections (3.95/1000 ds) and 34 episodes of clinical sepsis (3.53/1000 ds). Seventy-one vascular access infections without bacteremia were identified (7.38/1000 ds), including 16 permanent-fistulae or graft infections (2.94/1000 ds), 7 permanent-tunneled central catheter infections (3.93/1000 ds), and 48 temporary-catheter infections (19.92/1000 ds). Staphylococcal organisms were responsible for 77% of the infections, with Staphylococcus epidermidis being the strain most commonly implicated. Gram-negative organisms were responsible for 23% of the infections. In conclusion, infection rates were highest in hemodialysis patients with temporary vascular access, com-
I. Saeed (*) · S.H. Al-Mueilo · H.A. Bokhary · G.O.A. Ladipo · A. Al-Rubaish King Fahd Hospital of the University, P.O. Box 40032, Al Khobar 31952, Saudi Arabia Tel. ⫹966-3-8645928; Fax ⫹966-3-8822289 e-mail: [email protected] or [email protected]
pared with rates in those with permanent arteriovenous fistulae and synthetic grafts. Most of the bacterial organisms isolated from the vascular access sites were gram-positive cocci, with S. epidermidis accounting for 50% of the organisms. The rate of infection with gram-negative bacilli was higher than in other reports. Our greater dependence on central venous catheters, due to local factors, coupled with the immune-compromising comorbid conditions of our patients, may be contributory to the pattern of infection reported. Delays in the creation of vascular grafts for hemodialysis access should be avoided. Key words Vascular access · Infection · Hemodialysis · Bacteremia
Introduction Among the complications associated with vascular access in hemodialysis patients, infection is the principal cause of morbidity and the second leading cause of mortality.1,2 This is due to the repeated need to access the vascular system. Methods to provide longterm access to the vascular system of hemodialysis patients are well known. An arteriovenous (AV) fistula or graft can be surgically created to arterialize a segment of the venous system; a central venous catheter (CVC) may be inserted, usually in the subclavian, internal jugular, or femoral vein. The CVC may be temporary when used for immediate hemodialysis in acute situations, or permanent, with the advantage of avoiding repeated central venous cannulation. Strategies for preventing vascular access infections have been proposed by the Centers for Disease Control and Prevention (CDC).3 Surveillance of nosocomial or hospital-acquired infections is a key element of all infection control programs,4 and can be directly applied for hemodialysis vascular access infection. Hemodialysis by means of an AV fistula results in a lower rate of bacteremia than dialysis via an AV graft,5 whereas infection rates are higher for hemodialysis through CVCs than through AV fistulae and grafts.5,6
243
Most reports of access-related infections in hemodialysis patients have incriminated staphylococcus organisms as the leading cause of bacteremia, while gram-negative bacteremia was considered to be uncommon.17–19,21 We have recently experienced an increased rate of hemodialysisrelated gram-negative bacteremia at our institution and this motivated us to undertake this study.
Patients and methods Study design We prospectively studied all hemodialysis patients at King Fahd Hospital of the University, Al-Khobar, Saudi Arabia, for vascular access infection over an 18-month period (August 1999 to January 2001). The total number of dialysis sessions was recorded on a daily basis and grouped according to the vascular access type: temporary central venous catheter (CVC), permanent tunneled central catheter, and permanent AV fistula or graft. At each dialysis session, patients were evaluated by the medical and trained nursing staff for the presence of vascular access infection. Cultures were obtained from infected local vascular access sites, catheter tips, and blood of affected patients. The presence of infections was recorded and tabulated by a trained nurse under the supervision of a nephrologist.
Case definitions The CDC definitions for infections with or without bacteremia were used.7 Blood for culture was obtained by drawing blood from the dialysis arterial or venous line or port. Positive cultures reported by the clinical microbiology laboratory were reviewed daily by staff from the hospital’s infection control unit.
Microbiological tests Swabs from vascular access sites were inoculated onto agar media using standard techniques.9 Catheter tips were cultured according to a semiquantitative method.10 Blood cultures were performed for aerobic and anaerobic organisms
according to established methods.11 Bacteria were identified using the Microscan (Microscan system, Renton, WA, USA) system.11 Data analysis Infection rates were calculated by vascular access type for total, bloodstream, clinical sepsis, and vascular access site. Data analysis was performed according to the CDC National Infection Surveillance System.8 Statistical analysis was performed with Epi Info (Epi Info, 2000, USD, Snellville, USA).
Results Table 1 shows a breakdown of the patient population studied. A total of 183 patients had hemodialysis during the 18-month study period. The majority (86.3%) were Saudis; 57.4% were aged 50 years and above, and 55.2% were male. Table 2 shows the number of dialysis sessions according to the vascular access type. There were 9627 dialysis sessions, with a monthly mean of 540 ⫾ 78. A total of 109 vascular access infections were detected during the study period, with an incidence rate of 11.3 infections/1000 dialysis sessions (ds) (Table 3). The highest incidence rate was with temporary catheters (28.23 infections/ 1000 ds), while the lowest incidence rate was with fistulae and grafts (4.23/1000 ds); in the incidence rate was between these two levels with the tunneled central catheters (10.11/ 1000 ds). The infection rate of fistulae and grafts was significantly lower than the rates for the other two access types (P ⬍ 0.001), while that for tunneled catheters was significantly lower than that for temporary catheters (P ⬍ 0.01). Thirty-eight (34.9%) of the 109 vascular access infections were complicated by bacteremia. The bacteremia rate was highest with temporary central catheters (8.30/1000 ds) and lowest with fistulae and grafts (1.29/1000 ds). The bacteremia rate for fistulae and grafts was significantly lower than the rates for the temporary and tunneled central catheters (P ⬍ 0.001). Clinical sepsis occurred in 34 (31.19%) of the 109 infections. The clinical sepsis rate was significantly lower for fistulae and grafts than for the temporary and tunneled central catheters (P ⬍ 0.001). The rate of access-site infec-
Table 1. Breakdown of patients studied Patient nos. by ages in decades 10–19 Nationality Male Female
S NS 30 10
Total
40
S, Saudi; NS, non-Saudi
20–29
30–39
40–49
50–59
60–69
S NS 51 80
S NS 83 71
S NS 18 4 17 2
S NS 33 3 23 4
S NS 13 3 11 3
35 6
56 7
24 6
13 1
15 4
⭌70 S NS 61 50 11 1
Total S NS 86 15 72 10 158 25
244
tion without bacteremia was, similarly, lowest among patients with fistulae and grafts, with the same degree of statistical significance, compared with the rates for the temporary catheters.
Table 2. Number of dialysis sessions according to vascular access type Access type
No.
Temporary central catheter Tunneled central catheter Fistula/graft
2409 1781 5437
Total
9627
Table 3. Data summary of vascular access infections Characteristic Total number of vascular access infections Temp. CC Tunn. CC Fistulae/grafts Total Bacteremia Temp. CC Tunn. CC Fistulae/grafts Total Clinical sepsis episodes Temp. CC Tunn. CC Fistulae/grafts Total Access-site infections Temp. CC Tunn. CC Fistulae/grafts Total
No. (rate)a (28.23) 18 (10.11) 23 (4.23) 109 (11.32) 20 (8.30) 11 (6.18) 7 (1.29) 38 (3.95) 18 (7.47) 10 (5.61) 6 (1.10) 34 (3.53) 48 (19.92) 7 (3.93) 16 (2.94) 71 (7.38)
Temp. CC, Temporary central catheters; Tunn. CC, tunneled central catheters a Rates are expressed as the number of infections/1000 dialysis sessions
Fig. 1. Prevalence of infecting organisms. Staph. epid, Staphylococcus epidermidis; MRSA, methicillin-resistant Staphylococcus aureus; gm-ve bac, gram-negative bacilli
Table 4 shows the breakdown of findings for organisms isolated from access sites and blood, and the total number of organisms isolated. Seventy-seven percent of the bacterial organisms isolated from the blood and vascular access sites were gram-positive cocci, while 23% of the isolates were gram-negative bacilli. Figure 1 shows the prevalence of the infecting organisms. Staphylococcus epidermidis was causative in 49.5% of the infections, methicillin-resistant Staphylococcus aureus in 10.1%, and methicillin-sensitive S. aureus in 17.4%. All of the staphylococci in this study were vancomycin-sensitive. Pseudomonas aeruginosa, Klebsiella, and Enterobacter species were the predominant causative gram-negative bacilli. Removal of the temporary or tunneled central catheters was followed by eradication of infection in 92.7% of cases, while in the remaining 7.3%, infection was difficult to control because of the immunocompromised state of the patients. Associated or comorbid conditions in our patients were: diabetes mellitus (35 cases), systemic lupus erythematosus (10 cases), chronic liver disease (9 cases), patients on recent or longterm steroid therapy (10 cases), and multiple myeloma (3 cases).
Table 4. Breakdown of organisms isolated from access sites and blood, and the total number of organisms isolated Organisms
Access site no (%)
Blood no (%)
Staphylococcus epidermidis S. aureus MRSA Pseudomonas aeruginosa Klebsiella pneumoniae Enterobacter cloacae Escherichia coli Hemophilus influenzae
34 (47.9) 15 (21.1) 6 (8.5) 6 (8.5) 4 (5.6) 3 (4.2) 2 (2.8) 1 (1.4)
20 (52.6) 4 (10.5) 5 (13.2) 3 (7.9) 3 (7.9) 2 (5.3) 1 (2.6) 0 (0)
54 (49.5) 19 (17.4) 11 (10.1) 9 (8.3) 7 (6.4) 5 (4.6) 3 (2.8) 1 (0.9)
Total
71 (100)
38 (100)
109 (100)
MRSA, Methicillin-resistant S. aureus
59.6%
22% 10.1% 8.3%
Total no (%)
245
Discussion Bacteremic infections are the major cause of morbidity and mortality in chronic hemodialysis patients.12 Central venous catheters (CVCs) have become an indispensable form of hemodialysis access, but the incidence rate of infection with CVCs is double those of other access types combined. Our result is in agreement with a literature report noting that sepsis, as a consequence of central venous hemodialysis catheters, is the major cause of morbidity in the hemodialysis population.13 The majority of these catheters become colonized, with consequent peripheral bacteremia.13 In fact, patients on dialysis are at a high risk of Staphylococcus aureus bacteremia and they have a fourfold higher mortality from CVC-related S. aureus bacteremia than other patients.14 Dittmer et al.13 found that bacterial colonization of CVCs often led to bacteremia, and that the risk of subsequent bacteremia was related not only to time left in situ but also to the degree of colonization. They concluded that surveillance cultures would allow clinicians to detect colonization before bacteremia occurred and thus take preventive measures. On the other hand, Hoen et al.12 described four risk factors for bacteremia: (1) vascular access (catheter versus fistula: relative risk [RR], 7.6); (2) history of bacteremia (ⱖ2 episodes versus no previous episode: RR, 7.3); (3) immunosuppressive therapy (current versus none, RR, 3.0); and (4) corpuscular hemoglobin (1 g/dl increment: RR, 0.7). They also found that long-term implanted catheters were the leading risk factors for bacteremia in chronic hemodialysis patients and caused a trend toward recurrence of bacteremia that was not associated with chronic staphylococcal nasal carriage.12 Tunneled central catheters were thought to be safer in this respect; however, recent publications have shown that tunneled dialysis catheters are complicated by frequent systemic infections, and that the treatment of bacteremia associated with infected tunneled catheters requires both antibiotic therapy and catheter removal.15 The bacteremia rate for tunneled catheters in our patients was significantly higher than that for fistulae and grafts (P ⬍ 0.001), but lower than that for CVC. Many authors who have investigated access-related infections in hemodialysis patients have reported S. aureus organisms as the leading cause of bacteremia in this setting. Kurt et al.16 reported a prevalence of 82%, while other studies reported prevalences between 52% and 67%.17,18 Our finding differs from the reported pattern of infection, because more than half of the organisms isolated from blood were S. epidermidis rather than S. aureus. This may point to the emergence of more virulent strains among this group of organisms. Gram-negative bacteremia was considered to be uncommon among hemodialysis patients, and an 11% prevalence was considered as an outbreak of gram-negative bacteremia by some authors.20 This was attributed to the waste-handling option ports. The 22% rate of gramnegative bacteremia in our study after elimination of the waste-handling option ports is higher than previously quoted figures.17,29,40 Almost all of these gram-negative bacteremias in our study (98.6%) occurred in patients
dialysed with central catheters. This high rate of gramnegative infection may be attributed to the fact that many of our patients may have been immunocompromised by their comorbid conditions (such as systemic lupus erythematosus, multiple myeloma, diabetes mellitus, and chronic liver disease). Inadequate disinfection could also be a contributory factor. Our growing dependence on central venous hemodialysis catheters has to be addressed to curtail the unwarranted rise in the rate of gram-negative infections. Reported national infection rates for local and bacteremic infections are 1%–4% for native AV fistulae21,22 and 11%–20% for AV grafts.21,23–25 Reported infection rates for cuffed central catheters depend on the duration of use,26,27 with 5%–8% bacteremia rates reported for catheters in use less than 3 months and rates approaching 50% for catheters in use for 12 months or more.26 These values may discourage the routine use of CVCs. Our documented infection rates of total access infection were 4.2 per 1000 dialysis sessions (ds) for AV fistulae and grafts, 10.11/ 1000 ds for tunneled central catheters, and 28.23/1000 ds for temporary central catheters, with bacteremia rates of 8.3/ 1000 ds with temporary central catheters, 6.18/1000 ds with tunneled central catheters, and 1.29/1000 ds with fistulae/ grafts. Lack of hygienic precautions during home care of the CVCs may be contributory to these infections, although our figures compare favorably with others. Temporary non-tunneled subclavian hemodialysis catheters have been associated with a rate of bloodstream infections that exceeds that reported for all other types of subclavian catheters used in hemodialysis.27,28 Also, the use of these catheters may be associated with bacterial endocarditis and septic pulmonary emboli.29–32 This may be related to dressing changes by inadequately trained personnel, prolonged duration of catheterization, and a high number of hemodialysis runs.3 Other studies have suggested that permanent tunneled hemodialysis catheters were associated with significantly lower bacteremia rates than non-tunneled catheters.33,34 Although the rates of bacteremia and clinical sepsis in our study were higher for temporary hemodialysis catheters than for permanent tunneled catheters, these differences were not significant. Our rates of bacteremia (8.3/1000 ds) and clinical sepsis (7.47/1000 ds) for CVCs were much higher than those in recent reports, i.e., 3.22/1000 ds and 1.76/1000 ds, respectively.35 The higher rates, again, may be due to defects in dressing policies, prolonged duration of catheterization, and deficiencies in the patient’s own personal hygiene. Local factors in our region have often resulted in delays in the creation of vascular grafts for hemodialysis access and these delays have, in turn, increased our dependence on central venous hemodialysis catheters as an alternative access. Staphylococci have become the most frequently isolated pathogen associated with catheter-related infections.36,37 Recent studies have reported that 40%–60% of hemodialysis patients are carriers of S. aureus in their nares.38 In addition, recent epidemiologic studies have shown that most S. aureus isolates causing infections in hemodialysis patients are from the patient’s endogenous flora.39
246
Strategies for the prevention of vascular access infections have been proposed by the National Kidney Foundation Dialysis Outcome Quality Initiative and by the CDC.3 The calculation of infection rates, as described in this study, allows for the longitudinal tracking of incidence rates during the evaluation of specific interventions. Specific measures to reduce the infection rate in our hemodialysis population are currently being studied and will be considered in future reports.
References 1. Churchill DN, Taylor DW, Cook RJ, La Plante P, Paul B, Cartier Pl, et al. Canadian hemodialysis morbidity study. Am J Kidney Dis 1992;19:214–34. 2. Johnson WJ, Kurtz SB, Mitchell JC III, Van Den Berg CJ, Wochos DN, O’Fallon W, et al. Results of treatment of hemodialysis center patients. Mayo Clin Proc 1984;59:669–71. 3. Pearson ML. The Hospital Infection Control Practices Advisory Committee. Guidelines for prevention of intravascular devicerelated infections. Infect Control Hosp Epidemiol 1996;17:438–73. 4. Haley R, Culver DH, White JW. The efficacy of infection surveillance and control programs in preventing nosocomial infections in US hospitals. Am J Epidemiol 1985;121:182–205. 5. Hoen B, Kessler M, Hestin D, Mayeux D. Risk factors for bacterial infections in chronic hemodialysis adult patients: a multicenter prospective study. Nephrol Dial Transplant 1995;10:377–81. 6. Cheesbrough JS, Finch RG, Burden RP. A prospective study of the mechanism of infection associated with hemodialysis catheters. J Infect Dis 1986;154:579–89. 7. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:28–40. 8. Gaynes RP, Culver DH, Emori TG, Horan TC, Banerjee SN, Edward JR, et al. The National Nosocomial Infections Surveillance System: plans for the 1990s and beyond. Am J Med 1991;91 (Suppl 3B):116S–20S. 9. Cintron F. Initial processing, inoculation, and incubation of aerobic bacteriology specimens. In: Isenberg HD, editor. Clinical microbiology procedures handbook. Washington, DC; 1994. p. 1.4.1–1.4.19. 10. Maki DG, Weise CE, Sarafin HW. A semi-quantitative method for identifying intravenous catheter infection. N Engl J Med 1977;296:1305–9. 11. Almon R, Pezzlo M. Processing and interpretation of blood cultures. In: Isenberg HD, editor. Clinical microbiology procedures handbook. Washington, DC: American Society for Microbiology; 1994. p 1.7.1–1.7.11. 12. Hoen B, Paul-Dauphin A, Hestin D, Kessler M. A multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients. J Am Soc Nephrol 1998;9:869–76. 13. Dittmer ID, Sharp D, McNulty CA, Williams AJ, Banks RA. A prospective study of central venous hemodialysis catheter colonization and peripheral bacteremia. Clin Nephrol 1999;51:34–9. 14. Nielsen J, Kolmos HJ, Espersen F. Staphylococcus aureus bacteremia among patients undergoing dialysis – focus on dialysis catheter-related cases. Nephrol Dial Transplant 1998;13:139–45. 15. Tanriover B, Carlton D, Saddekni S, Hamrick K, Oser R, Westfall AO, et al. Bacteremia associated with tunneled dialysis catheters: comparison of two treatment strategies. Kidney Int 2000;57:2151– 5. 16. Stevenson KB, Adcox MJ, Mallea MC, Nagraj N, Jon PW. Standardized surveillance of hemodialysis vascular access infections: 18-month experience at an outpatient multifacility hemodialysis center. Infect Control Hosp Epidemiol 2000;21:200–3. 17. Taylor GD, Margaret MK, Maureen BC, Louise C, Linda CH, Kinga KG. Central venous catheters as a source of hemodialysisrelated bacteremia. Infect Control Hosp Epidemiol 1998;19:643–6.
18. Rodriguez GA, Carton JA, Lopez PB, Casado J, Perez F, Aguado S. Bacteremia in patients undergoing chronic hemodialysis in a 16year period. Rev Clin Esp 1997;197:484–9. 19. Wang SA, Levine RB, Carson LA, Arduino MJ, Killar T, Grillo FG, et al. An outbreak of gram-negative bacteremia in hemodialysis patients traced to hemodialysis machine waste drain ports. Infect Control Hosp Epidemiol 1999;20:746–51. 20. Kherlakian GM, Roedersheimer LR, Arbaugh JJ, Newmark KJ, King LR. Comparison of autogenous fistula versus expanded polytetrafluoroethylene graft fistula for angio-access in hemodialysis. Am J Surg 1986;152:238–43. 21. Nazzal MM, Neglern P, Naseem J, Christenson JT, Al Hassan HK. The brachiocephalic fistula: a successful secondary vascular access procedure. Vasa 1990;19:326–9. 22. Dunlop MG, Mackilay JY, Jenkins AM. Vascular access experience with the brachiocephalic fistula. Ann R Coll Surg Engl 1986;68:203–6. 23. Raju S. PTFE grafts for hemodialysis: techniques for insertion and management of complications. Ann Surg 1987;206:666–73. 24. Odurny A, Slopak M. The use of Goretex (P.T.E.F.) for angioaccess for chronic hemodialysis: the place of peri-operative antibiotics. Br J Clin Pract 1984;38:134–7. 25. Suchoki P, Conlon PJ, Krelson M, Harland RC, Schwab SJ. Silastic cuffed catheters for hemodialysis vascular access: thrombolytic and mechanical correction of HD catheters malfunction. Am J Kidney Dis 1996;28:379–86. 26. Vanholder R, Hoenich N, Ringoir S. Morbidity and mortality of central venous catheter hemodialysis: a review of 10 years experience. Nephron 1987;47:274–9. 27. Blake PG, Huraib S, Vldall PR. The use of the dual lumen jugular venous catheters as definitive long term access for hemodialysis. Int Artif Organs 1990;13:26–31. 28. Pezzarossi HF, Ponce de Leon RS, Calva JJ, Lazo de la Vega SA, Ruiz Palacios GM. High incidence of subclavian catheter-related bacteremia. Infect Control Hosp Epidemiol 1986;7:596–9. 29. Keane WF, Shapiro FL, Raij L. Incidence and type of infection occurring in 445 chronic hemodialysis patients. Trans Am Soc Artif Intern Organs 1977;23:42–7. 30. Shaffer D, Madras PN, Williams ME, D’Elia JA, Kaladany A, Monaco AP. Use of Dacron cuffed silicon catheters as long-term access. ASAIO J 1992;38:55–8. 31. Robinson DL, Fowler VG, Sexton DJ, Corey RG, Conlon PJ. Bacterial endocarditis in hemodialysis patients. Am J Kidney Dis 1997;30:521–42. 32. Schwab SJ, Buller GL, Mc Cann RL, Bollinger RR, Stickel DL. Prospective evaluation of a Dacron-cuffed hemodialysis catheter for prolonged use. Am J Kidney Dis 1988;11:166–9. 33. Vanherwenghem JL, Cabolet P, Dhaene M. Complications related to subclavian catheters for hemodialysis. Int J Artif Organs 1986;5:297–309. 34. Bonormo RA, Rice D, Whalen C, Linn D, Ekstein E, Shales DM. Risk factors associated with permanent access-site infections in chronic hemodialysis patients. Infect Control Hosp Epidemiol 1997;18:757–61. 35. Banerjee SN, Emori TG, Culver HD. Secular trends in noscomial primary blood stream infections in the United States, 1980–1989. National Noscomial Infections Surveillance System. Am J Med 1991;91 (Suppl 3B):86S–9S. 36. Schaberg DR, Culver DH, Gayness RP. Major trends in the microbial etiology of nosocomial infection. Am J Med 1991;91 (Suppl 3B):72S–5S. 37. Kirmani N, Tuazon CU, Murray HW, Parrish AE, Sheagren JN. Staphylococcus aureus carriage rate of patients receiving long-term hemodialysis. Arch Intern Med 1978;138:1657–9. 38. Ena J, Boelaert JR, Boyken BA, Van Landuyt HW, Godard CA, Herwaldt LA. Epidemiology of Staphylococcus aureus infections in patients on hemodialysis. Infect Control Hosp Epidemiol 1994;15:78–81. 39. National Kidney Foundation. Dialysis outcomes quality initiative. Clinical practice guidelines. Am J Kidney Dis 1997;30(Suppl): S137–S240. 40. Kairaitis LK, Gottlieb T. Outcome and complications of temporary hemodialysis catheters. Nephrol Dial Transplant 1999;14:1710– 4.
© Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases 2002
ORIGINAL ARTICLE Ibrahiem Saeed Abdulrahman · Samir H. Al-Mueilo Huda A. Bokhary · Gani O.A. Ladipo Abdulla Al-Rubaish
A prospective study of hemodialysis access-related bacterial infections
Received: November 8, 2001 / Accepted: April 25, 2002
Abstract The objective of this study was to describe hemodialysis vascular-access related infections that occurred in hemodialysis patients over an 18-month period. The study is a prospective descriptive analysis of incidence infection rates in a hemodialysis unit in a tertiary-care medical center. Prospective surveillance for hemodialysis vascular accessrelated infection was performed for all patients undergoing hemodialysis from November 1999 through April 2001 at King Fahd Hospital of King Faisal University, Al-Khobar, Saudi Arabia. The total number of dialysis sessions was calculated. The type of vascular access was noted. Cultures were obtained and all infections were recorded and infection rates were calculated. There were 9627 hemodialysis sessions (5437 via permanent fistulae or grafts, 2409 via temporary central catheters, and 1781 via permanent tunneled catheters) during the 18-month study period. We identified a total of 109 infections, for a rate of 11.32/1000 dialysis sessions (ds). Of the 109, 23 involved permanent fistulae or grafts (4.23/1000 ds); 18 involved permanenttunneled central catheter infections (10.1/1000 ds); and 68 involved temporary-catheter infections (28.23/1000 ds). There were 38 bloodstream infections (3.95/1000 ds) and 34 episodes of clinical sepsis (3.53/1000 ds). Seventy-one vascular access infections without bacteremia were identified (7.38/1000 ds), including 16 permanent-fistulae or graft infections (2.94/1000 ds), 7 permanent-tunneled central catheter infections (3.93/1000 ds), and 48 temporary-catheter infections (19.92/1000 ds). Staphylococcal organisms were responsible for 77% of the infections, with Staphylococcus epidermidis being the strain most commonly implicated. Gram-negative organisms were responsible for 23% of the infections. In conclusion, infection rates were highest in hemodialysis patients with temporary vascular access, com-
I. Saeed (*) · S.H. Al-Mueilo · H.A. Bokhary · G.O.A. Ladipo · A. Al-Rubaish King Fahd Hospital of the University, P.O. Box 40032, Al Khobar 31952, Saudi Arabia Tel. ⫹966-3-8645928; Fax ⫹966-3-8822289 e-mail: [email protected] or [email protected]
pared with rates in those with permanent arteriovenous fistulae and synthetic grafts. Most of the bacterial organisms isolated from the vascular access sites were gram-positive cocci, with S. epidermidis accounting for 50% of the organisms. The rate of infection with gram-negative bacilli was higher than in other reports. Our greater dependence on central venous catheters, due to local factors, coupled with the immune-compromising comorbid conditions of our patients, may be contributory to the pattern of infection reported. Delays in the creation of vascular grafts for hemodialysis access should be avoided. Key words Vascular access · Infection · Hemodialysis · Bacteremia
Introduction Among the complications associated with vascular access in hemodialysis patients, infection is the principal cause of morbidity and the second leading cause of mortality.1,2 This is due to the repeated need to access the vascular system. Methods to provide longterm access to the vascular system of hemodialysis patients are well known. An arteriovenous (AV) fistula or graft can be surgically created to arterialize a segment of the venous system; a central venous catheter (CVC) may be inserted, usually in the subclavian, internal jugular, or femoral vein. The CVC may be temporary when used for immediate hemodialysis in acute situations, or permanent, with the advantage of avoiding repeated central venous cannulation. Strategies for preventing vascular access infections have been proposed by the Centers for Disease Control and Prevention (CDC).3 Surveillance of nosocomial or hospital-acquired infections is a key element of all infection control programs,4 and can be directly applied for hemodialysis vascular access infection. Hemodialysis by means of an AV fistula results in a lower rate of bacteremia than dialysis via an AV graft,5 whereas infection rates are higher for hemodialysis through CVCs than through AV fistulae and grafts.5,6
243
Most reports of access-related infections in hemodialysis patients have incriminated staphylococcus organisms as the leading cause of bacteremia, while gram-negative bacteremia was considered to be uncommon.17–19,21 We have recently experienced an increased rate of hemodialysisrelated gram-negative bacteremia at our institution and this motivated us to undertake this study.
Patients and methods Study design We prospectively studied all hemodialysis patients at King Fahd Hospital of the University, Al-Khobar, Saudi Arabia, for vascular access infection over an 18-month period (August 1999 to January 2001). The total number of dialysis sessions was recorded on a daily basis and grouped according to the vascular access type: temporary central venous catheter (CVC), permanent tunneled central catheter, and permanent AV fistula or graft. At each dialysis session, patients were evaluated by the medical and trained nursing staff for the presence of vascular access infection. Cultures were obtained from infected local vascular access sites, catheter tips, and blood of affected patients. The presence of infections was recorded and tabulated by a trained nurse under the supervision of a nephrologist.
Case definitions The CDC definitions for infections with or without bacteremia were used.7 Blood for culture was obtained by drawing blood from the dialysis arterial or venous line or port. Positive cultures reported by the clinical microbiology laboratory were reviewed daily by staff from the hospital’s infection control unit.
Microbiological tests Swabs from vascular access sites were inoculated onto agar media using standard techniques.9 Catheter tips were cultured according to a semiquantitative method.10 Blood cultures were performed for aerobic and anaerobic organisms
according to established methods.11 Bacteria were identified using the Microscan (Microscan system, Renton, WA, USA) system.11 Data analysis Infection rates were calculated by vascular access type for total, bloodstream, clinical sepsis, and vascular access site. Data analysis was performed according to the CDC National Infection Surveillance System.8 Statistical analysis was performed with Epi Info (Epi Info, 2000, USD, Snellville, USA).
Results Table 1 shows a breakdown of the patient population studied. A total of 183 patients had hemodialysis during the 18-month study period. The majority (86.3%) were Saudis; 57.4% were aged 50 years and above, and 55.2% were male. Table 2 shows the number of dialysis sessions according to the vascular access type. There were 9627 dialysis sessions, with a monthly mean of 540 ⫾ 78. A total of 109 vascular access infections were detected during the study period, with an incidence rate of 11.3 infections/1000 dialysis sessions (ds) (Table 3). The highest incidence rate was with temporary catheters (28.23 infections/ 1000 ds), while the lowest incidence rate was with fistulae and grafts (4.23/1000 ds); in the incidence rate was between these two levels with the tunneled central catheters (10.11/ 1000 ds). The infection rate of fistulae and grafts was significantly lower than the rates for the other two access types (P ⬍ 0.001), while that for tunneled catheters was significantly lower than that for temporary catheters (P ⬍ 0.01). Thirty-eight (34.9%) of the 109 vascular access infections were complicated by bacteremia. The bacteremia rate was highest with temporary central catheters (8.30/1000 ds) and lowest with fistulae and grafts (1.29/1000 ds). The bacteremia rate for fistulae and grafts was significantly lower than the rates for the temporary and tunneled central catheters (P ⬍ 0.001). Clinical sepsis occurred in 34 (31.19%) of the 109 infections. The clinical sepsis rate was significantly lower for fistulae and grafts than for the temporary and tunneled central catheters (P ⬍ 0.001). The rate of access-site infec-
Table 1. Breakdown of patients studied Patient nos. by ages in decades 10–19 Nationality Male Female
S NS 30 10
Total
40
S, Saudi; NS, non-Saudi
20–29
30–39
40–49
50–59
60–69
S NS 51 80
S NS 83 71
S NS 18 4 17 2
S NS 33 3 23 4
S NS 13 3 11 3
35 6
56 7
24 6
13 1
15 4
⭌70 S NS 61 50 11 1
Total S NS 86 15 72 10 158 25
244
tion without bacteremia was, similarly, lowest among patients with fistulae and grafts, with the same degree of statistical significance, compared with the rates for the temporary catheters.
Table 2. Number of dialysis sessions according to vascular access type Access type
No.
Temporary central catheter Tunneled central catheter Fistula/graft
2409 1781 5437
Total
9627
Table 3. Data summary of vascular access infections Characteristic Total number of vascular access infections Temp. CC Tunn. CC Fistulae/grafts Total Bacteremia Temp. CC Tunn. CC Fistulae/grafts Total Clinical sepsis episodes Temp. CC Tunn. CC Fistulae/grafts Total Access-site infections Temp. CC Tunn. CC Fistulae/grafts Total
No. (rate)a (28.23) 18 (10.11) 23 (4.23) 109 (11.32) 20 (8.30) 11 (6.18) 7 (1.29) 38 (3.95) 18 (7.47) 10 (5.61) 6 (1.10) 34 (3.53) 48 (19.92) 7 (3.93) 16 (2.94) 71 (7.38)
Temp. CC, Temporary central catheters; Tunn. CC, tunneled central catheters a Rates are expressed as the number of infections/1000 dialysis sessions
Fig. 1. Prevalence of infecting organisms. Staph. epid, Staphylococcus epidermidis; MRSA, methicillin-resistant Staphylococcus aureus; gm-ve bac, gram-negative bacilli
Table 4 shows the breakdown of findings for organisms isolated from access sites and blood, and the total number of organisms isolated. Seventy-seven percent of the bacterial organisms isolated from the blood and vascular access sites were gram-positive cocci, while 23% of the isolates were gram-negative bacilli. Figure 1 shows the prevalence of the infecting organisms. Staphylococcus epidermidis was causative in 49.5% of the infections, methicillin-resistant Staphylococcus aureus in 10.1%, and methicillin-sensitive S. aureus in 17.4%. All of the staphylococci in this study were vancomycin-sensitive. Pseudomonas aeruginosa, Klebsiella, and Enterobacter species were the predominant causative gram-negative bacilli. Removal of the temporary or tunneled central catheters was followed by eradication of infection in 92.7% of cases, while in the remaining 7.3%, infection was difficult to control because of the immunocompromised state of the patients. Associated or comorbid conditions in our patients were: diabetes mellitus (35 cases), systemic lupus erythematosus (10 cases), chronic liver disease (9 cases), patients on recent or longterm steroid therapy (10 cases), and multiple myeloma (3 cases).
Table 4. Breakdown of organisms isolated from access sites and blood, and the total number of organisms isolated Organisms
Access site no (%)
Blood no (%)
Staphylococcus epidermidis S. aureus MRSA Pseudomonas aeruginosa Klebsiella pneumoniae Enterobacter cloacae Escherichia coli Hemophilus influenzae
34 (47.9) 15 (21.1) 6 (8.5) 6 (8.5) 4 (5.6) 3 (4.2) 2 (2.8) 1 (1.4)
20 (52.6) 4 (10.5) 5 (13.2) 3 (7.9) 3 (7.9) 2 (5.3) 1 (2.6) 0 (0)
54 (49.5) 19 (17.4) 11 (10.1) 9 (8.3) 7 (6.4) 5 (4.6) 3 (2.8) 1 (0.9)
Total
71 (100)
38 (100)
109 (100)
MRSA, Methicillin-resistant S. aureus
59.6%
22% 10.1% 8.3%
Total no (%)
245
Discussion Bacteremic infections are the major cause of morbidity and mortality in chronic hemodialysis patients.12 Central venous catheters (CVCs) have become an indispensable form of hemodialysis access, but the incidence rate of infection with CVCs is double those of other access types combined. Our result is in agreement with a literature report noting that sepsis, as a consequence of central venous hemodialysis catheters, is the major cause of morbidity in the hemodialysis population.13 The majority of these catheters become colonized, with consequent peripheral bacteremia.13 In fact, patients on dialysis are at a high risk of Staphylococcus aureus bacteremia and they have a fourfold higher mortality from CVC-related S. aureus bacteremia than other patients.14 Dittmer et al.13 found that bacterial colonization of CVCs often led to bacteremia, and that the risk of subsequent bacteremia was related not only to time left in situ but also to the degree of colonization. They concluded that surveillance cultures would allow clinicians to detect colonization before bacteremia occurred and thus take preventive measures. On the other hand, Hoen et al.12 described four risk factors for bacteremia: (1) vascular access (catheter versus fistula: relative risk [RR], 7.6); (2) history of bacteremia (ⱖ2 episodes versus no previous episode: RR, 7.3); (3) immunosuppressive therapy (current versus none, RR, 3.0); and (4) corpuscular hemoglobin (1 g/dl increment: RR, 0.7). They also found that long-term implanted catheters were the leading risk factors for bacteremia in chronic hemodialysis patients and caused a trend toward recurrence of bacteremia that was not associated with chronic staphylococcal nasal carriage.12 Tunneled central catheters were thought to be safer in this respect; however, recent publications have shown that tunneled dialysis catheters are complicated by frequent systemic infections, and that the treatment of bacteremia associated with infected tunneled catheters requires both antibiotic therapy and catheter removal.15 The bacteremia rate for tunneled catheters in our patients was significantly higher than that for fistulae and grafts (P ⬍ 0.001), but lower than that for CVC. Many authors who have investigated access-related infections in hemodialysis patients have reported S. aureus organisms as the leading cause of bacteremia in this setting. Kurt et al.16 reported a prevalence of 82%, while other studies reported prevalences between 52% and 67%.17,18 Our finding differs from the reported pattern of infection, because more than half of the organisms isolated from blood were S. epidermidis rather than S. aureus. This may point to the emergence of more virulent strains among this group of organisms. Gram-negative bacteremia was considered to be uncommon among hemodialysis patients, and an 11% prevalence was considered as an outbreak of gram-negative bacteremia by some authors.20 This was attributed to the waste-handling option ports. The 22% rate of gramnegative bacteremia in our study after elimination of the waste-handling option ports is higher than previously quoted figures.17,29,40 Almost all of these gram-negative bacteremias in our study (98.6%) occurred in patients
dialysed with central catheters. This high rate of gramnegative infection may be attributed to the fact that many of our patients may have been immunocompromised by their comorbid conditions (such as systemic lupus erythematosus, multiple myeloma, diabetes mellitus, and chronic liver disease). Inadequate disinfection could also be a contributory factor. Our growing dependence on central venous hemodialysis catheters has to be addressed to curtail the unwarranted rise in the rate of gram-negative infections. Reported national infection rates for local and bacteremic infections are 1%–4% for native AV fistulae21,22 and 11%–20% for AV grafts.21,23–25 Reported infection rates for cuffed central catheters depend on the duration of use,26,27 with 5%–8% bacteremia rates reported for catheters in use less than 3 months and rates approaching 50% for catheters in use for 12 months or more.26 These values may discourage the routine use of CVCs. Our documented infection rates of total access infection were 4.2 per 1000 dialysis sessions (ds) for AV fistulae and grafts, 10.11/ 1000 ds for tunneled central catheters, and 28.23/1000 ds for temporary central catheters, with bacteremia rates of 8.3/ 1000 ds with temporary central catheters, 6.18/1000 ds with tunneled central catheters, and 1.29/1000 ds with fistulae/ grafts. Lack of hygienic precautions during home care of the CVCs may be contributory to these infections, although our figures compare favorably with others. Temporary non-tunneled subclavian hemodialysis catheters have been associated with a rate of bloodstream infections that exceeds that reported for all other types of subclavian catheters used in hemodialysis.27,28 Also, the use of these catheters may be associated with bacterial endocarditis and septic pulmonary emboli.29–32 This may be related to dressing changes by inadequately trained personnel, prolonged duration of catheterization, and a high number of hemodialysis runs.3 Other studies have suggested that permanent tunneled hemodialysis catheters were associated with significantly lower bacteremia rates than non-tunneled catheters.33,34 Although the rates of bacteremia and clinical sepsis in our study were higher for temporary hemodialysis catheters than for permanent tunneled catheters, these differences were not significant. Our rates of bacteremia (8.3/1000 ds) and clinical sepsis (7.47/1000 ds) for CVCs were much higher than those in recent reports, i.e., 3.22/1000 ds and 1.76/1000 ds, respectively.35 The higher rates, again, may be due to defects in dressing policies, prolonged duration of catheterization, and deficiencies in the patient’s own personal hygiene. Local factors in our region have often resulted in delays in the creation of vascular grafts for hemodialysis access and these delays have, in turn, increased our dependence on central venous hemodialysis catheters as an alternative access. Staphylococci have become the most frequently isolated pathogen associated with catheter-related infections.36,37 Recent studies have reported that 40%–60% of hemodialysis patients are carriers of S. aureus in their nares.38 In addition, recent epidemiologic studies have shown that most S. aureus isolates causing infections in hemodialysis patients are from the patient’s endogenous flora.39
246
Strategies for the prevention of vascular access infections have been proposed by the National Kidney Foundation Dialysis Outcome Quality Initiative and by the CDC.3 The calculation of infection rates, as described in this study, allows for the longitudinal tracking of incidence rates during the evaluation of specific interventions. Specific measures to reduce the infection rate in our hemodialysis population are currently being studied and will be considered in future reports.
References 1. Churchill DN, Taylor DW, Cook RJ, La Plante P, Paul B, Cartier Pl, et al. Canadian hemodialysis morbidity study. Am J Kidney Dis 1992;19:214–34. 2. Johnson WJ, Kurtz SB, Mitchell JC III, Van Den Berg CJ, Wochos DN, O’Fallon W, et al. Results of treatment of hemodialysis center patients. Mayo Clin Proc 1984;59:669–71. 3. Pearson ML. The Hospital Infection Control Practices Advisory Committee. Guidelines for prevention of intravascular devicerelated infections. Infect Control Hosp Epidemiol 1996;17:438–73. 4. Haley R, Culver DH, White JW. The efficacy of infection surveillance and control programs in preventing nosocomial infections in US hospitals. Am J Epidemiol 1985;121:182–205. 5. Hoen B, Kessler M, Hestin D, Mayeux D. Risk factors for bacterial infections in chronic hemodialysis adult patients: a multicenter prospective study. Nephrol Dial Transplant 1995;10:377–81. 6. Cheesbrough JS, Finch RG, Burden RP. A prospective study of the mechanism of infection associated with hemodialysis catheters. J Infect Dis 1986;154:579–89. 7. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:28–40. 8. Gaynes RP, Culver DH, Emori TG, Horan TC, Banerjee SN, Edward JR, et al. The National Nosocomial Infections Surveillance System: plans for the 1990s and beyond. Am J Med 1991;91 (Suppl 3B):116S–20S. 9. Cintron F. Initial processing, inoculation, and incubation of aerobic bacteriology specimens. In: Isenberg HD, editor. Clinical microbiology procedures handbook. Washington, DC; 1994. p. 1.4.1–1.4.19. 10. Maki DG, Weise CE, Sarafin HW. A semi-quantitative method for identifying intravenous catheter infection. N Engl J Med 1977;296:1305–9. 11. Almon R, Pezzlo M. Processing and interpretation of blood cultures. In: Isenberg HD, editor. Clinical microbiology procedures handbook. Washington, DC: American Society for Microbiology; 1994. p 1.7.1–1.7.11. 12. Hoen B, Paul-Dauphin A, Hestin D, Kessler M. A multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients. J Am Soc Nephrol 1998;9:869–76. 13. Dittmer ID, Sharp D, McNulty CA, Williams AJ, Banks RA. A prospective study of central venous hemodialysis catheter colonization and peripheral bacteremia. Clin Nephrol 1999;51:34–9. 14. Nielsen J, Kolmos HJ, Espersen F. Staphylococcus aureus bacteremia among patients undergoing dialysis – focus on dialysis catheter-related cases. Nephrol Dial Transplant 1998;13:139–45. 15. Tanriover B, Carlton D, Saddekni S, Hamrick K, Oser R, Westfall AO, et al. Bacteremia associated with tunneled dialysis catheters: comparison of two treatment strategies. Kidney Int 2000;57:2151– 5. 16. Stevenson KB, Adcox MJ, Mallea MC, Nagraj N, Jon PW. Standardized surveillance of hemodialysis vascular access infections: 18-month experience at an outpatient multifacility hemodialysis center. Infect Control Hosp Epidemiol 2000;21:200–3. 17. Taylor GD, Margaret MK, Maureen BC, Louise C, Linda CH, Kinga KG. Central venous catheters as a source of hemodialysisrelated bacteremia. Infect Control Hosp Epidemiol 1998;19:643–6.
18. Rodriguez GA, Carton JA, Lopez PB, Casado J, Perez F, Aguado S. Bacteremia in patients undergoing chronic hemodialysis in a 16year period. Rev Clin Esp 1997;197:484–9. 19. Wang SA, Levine RB, Carson LA, Arduino MJ, Killar T, Grillo FG, et al. An outbreak of gram-negative bacteremia in hemodialysis patients traced to hemodialysis machine waste drain ports. Infect Control Hosp Epidemiol 1999;20:746–51. 20. Kherlakian GM, Roedersheimer LR, Arbaugh JJ, Newmark KJ, King LR. Comparison of autogenous fistula versus expanded polytetrafluoroethylene graft fistula for angio-access in hemodialysis. Am J Surg 1986;152:238–43. 21. Nazzal MM, Neglern P, Naseem J, Christenson JT, Al Hassan HK. The brachiocephalic fistula: a successful secondary vascular access procedure. Vasa 1990;19:326–9. 22. Dunlop MG, Mackilay JY, Jenkins AM. Vascular access experience with the brachiocephalic fistula. Ann R Coll Surg Engl 1986;68:203–6. 23. Raju S. PTFE grafts for hemodialysis: techniques for insertion and management of complications. Ann Surg 1987;206:666–73. 24. Odurny A, Slopak M. The use of Goretex (P.T.E.F.) for angioaccess for chronic hemodialysis: the place of peri-operative antibiotics. Br J Clin Pract 1984;38:134–7. 25. Suchoki P, Conlon PJ, Krelson M, Harland RC, Schwab SJ. Silastic cuffed catheters for hemodialysis vascular access: thrombolytic and mechanical correction of HD catheters malfunction. Am J Kidney Dis 1996;28:379–86. 26. Vanholder R, Hoenich N, Ringoir S. Morbidity and mortality of central venous catheter hemodialysis: a review of 10 years experience. Nephron 1987;47:274–9. 27. Blake PG, Huraib S, Vldall PR. The use of the dual lumen jugular venous catheters as definitive long term access for hemodialysis. Int Artif Organs 1990;13:26–31. 28. Pezzarossi HF, Ponce de Leon RS, Calva JJ, Lazo de la Vega SA, Ruiz Palacios GM. High incidence of subclavian catheter-related bacteremia. Infect Control Hosp Epidemiol 1986;7:596–9. 29. Keane WF, Shapiro FL, Raij L. Incidence and type of infection occurring in 445 chronic hemodialysis patients. Trans Am Soc Artif Intern Organs 1977;23:42–7. 30. Shaffer D, Madras PN, Williams ME, D’Elia JA, Kaladany A, Monaco AP. Use of Dacron cuffed silicon catheters as long-term access. ASAIO J 1992;38:55–8. 31. Robinson DL, Fowler VG, Sexton DJ, Corey RG, Conlon PJ. Bacterial endocarditis in hemodialysis patients. Am J Kidney Dis 1997;30:521–42. 32. Schwab SJ, Buller GL, Mc Cann RL, Bollinger RR, Stickel DL. Prospective evaluation of a Dacron-cuffed hemodialysis catheter for prolonged use. Am J Kidney Dis 1988;11:166–9. 33. Vanherwenghem JL, Cabolet P, Dhaene M. Complications related to subclavian catheters for hemodialysis. Int J Artif Organs 1986;5:297–309. 34. Bonormo RA, Rice D, Whalen C, Linn D, Ekstein E, Shales DM. Risk factors associated with permanent access-site infections in chronic hemodialysis patients. Infect Control Hosp Epidemiol 1997;18:757–61. 35. Banerjee SN, Emori TG, Culver HD. Secular trends in noscomial primary blood stream infections in the United States, 1980–1989. National Noscomial Infections Surveillance System. Am J Med 1991;91 (Suppl 3B):86S–9S. 36. Schaberg DR, Culver DH, Gayness RP. Major trends in the microbial etiology of nosocomial infection. Am J Med 1991;91 (Suppl 3B):72S–5S. 37. Kirmani N, Tuazon CU, Murray HW, Parrish AE, Sheagren JN. Staphylococcus aureus carriage rate of patients receiving long-term hemodialysis. Arch Intern Med 1978;138:1657–9. 38. Ena J, Boelaert JR, Boyken BA, Van Landuyt HW, Godard CA, Herwaldt LA. Epidemiology of Staphylococcus aureus infections in patients on hemodialysis. Infect Control Hosp Epidemiol 1994;15:78–81. 39. National Kidney Foundation. Dialysis outcomes quality initiative. Clinical practice guidelines. Am J Kidney Dis 1997;30(Suppl): S137–S240. 40. Kairaitis LK, Gottlieb T. Outcome and complications of temporary hemodialysis catheters. Nephrol Dial Transplant 1999;14:1710– 4.