Bloodstream infections in hemodialysis patients: getting some deserved attention.

Abstract

Health care–associated bloodstream infections (BSIs) are a major public health problem. Most surveillance and prevention efforts have been focused on hospital intensive care units.1 However, hemodialysis patients have long been known to be at high risk for infection and antimicrobial resistance. Because the incidence of end-stage renal disease doubled in the past decade, the number of hemodialysis patients at risk for BSI and other infections has increased rapidly. At the end of 2000, approximately 240,000 dialysis patients were treated in 3,700 outpatient facilities in the United States.2 This issue of Infection Control and Hospital Epidemiology contains three studies that focus some deserved attention on surveillance and prevention of BSI in hemodialysis patients.3-5 A related article discusses catheter-associated BSI in hospital inpatients.6 Hemodialysis patients require a vascular access site to withdraw blood so that waste substances can be removed and to replace the blood afterwards. These vascular access sites, listed in order of increasing infection risk, may be native arteriovenous fistulae, synthetic arteriovenous grafts, “permanent” (tunneled, cuffed) catheters, and “temporary” (nontunneled, noncuffed) catheters. Recently, an implanted port device has become available, and it is hoped that this device will have lower associated infection rates than hemodialysis catheters. Taylor et al.3 provide the results of prospective surveillance conducted in 11 Canadian dialysis units for 6 months during 1998 to 1999. The BSI definition that was used is substantially different from definitions used in the United States. Among other features, this definition would include episodes in which the same organism was isolated from blood and from the intravascular device or skin surface; and single positive cultures for skin contaminant organisms in patients who were immunocompromised (with this term remaining undefined in the article). There were 184 BSIs with an overall BSI rate of 0.59 BSI per 1,000 patient-days (to facilitate comparisons, all BSI rates presented here were converted to BSIs per 1,000 catheterdays; Table). Substantial variation in rates among the 11 centers was noted. In a second surveillance study, Dopirak et al.5 determined rates of BSI among patients at 10 dialysis centers in Connecticut during 1999 to 2000. The Centers for Disease Control and Prevention (CDC) definition for primary BSI was used. A total of 158 BSIs were noted with an overall rate of 0.45 per 1,000 patient-days (Table). Interestingly, rates declined sharply during the course of the study. Only a few years ago there were few reports of BSI surveillance among hemodialysis patients. In addition to the two studies reported in this issue, two other surveillance studies have recently been reported: one among 6 units in Idaho and Oregon,7 and the other among 109 centers participating in a system sponsored by the CDC.8 Despite differing definitions and surveillance methods, the reported BSI rates from these 4 surveillance studies are remarkably similar, with 3 of the 4 rounding to 0.6 BSI per 1,000 patient-days overall (Table). This similarity would suggest that adequate results can be obtained with a variety of surveillance methods—the important thing is to do surveillance and use the results to reduce infection rates. The method used to measure the denominator can make a substantial difference in the amount of work required for BSI surveillance. In studies of hemodialysis patients, BSI rates have been expressed as infections per 1,000 patient-days, 1,000 dialysis sessions3,5,7 (patients gen-