Gram-Negative Bloodstream Infections in Hemodialysis Outpatient Centers

In the United States, more than 450,000 patients receive hemodialysis in approximately 6500 outpatient centers. Patients receiving maintenance dialysis are at high risk for morbidity and mortality. In 2014, the Centers for Disease Control and Prevention (CDC) received 29,516 reports of bloodstream infections among hemodialysis outpatients.

The most common cause of bloodstream infections in hemodialysis patients is Gram-positive organisms; bloodstream infections due to Gram-negative organisms are less common. However, there have been reports of outbreaks associated with Gram-negative organisms in outpatient hemodialysis facilities attributed to water sources, such as contaminated reprocessed dialyzers, improper handling of medications, hemodialysis equipment, and dialysate. Infections associated with water reservoirs have also been reported.

Shannon A. Novosad, MD, MPH, and colleagues conducted matched case-control investigations at three outpatient hemodialysis facilities to examine an outbreak of Gram-negative bloodstream infections. Results of the investigations were reported in the American Journal of Kidney Diseases [2019;74(5):610-619].

Through an August 2016 review of routine surveillance data reported to the National Healthcare Safety Network, the CDC detected a cluster of five bloodstream infections caused by Serratia marcescens in an outpatient hemodialysis facility (Facility A). In subsequent consultations with state health departments, two additional outbreaks caused by similar Gram-negative organisms were identified at facilities owned by the same company (Facility B and Facility C). The most commonly involved Gram-negative organisms were S marcescens, Pseudomonas aeruginosa, and Enterobacter cloacae.

The investigation defined a case as a blood culture from which any Gram-negative bacteria was identified from a patient who received hemodialysis at Facility A, B, or C from July 1, 2105, to November 30, 2016. For more than one case in a single patient, there had to have been 21 days between positive cultures.

To examine risk factors for becoming a case, two 1:1-matched case-control investigations were performed at Facilities A and B. The focus of the first investigation was patient-specific risk factors such as age and comorbid conditions. Case-patients were matched with randomly selected control patients who received hemodialysis at the same facility and did not develop a Gram-negative bloodstream infection during the study period.

The exposures of interest were infection control practices, sources of water, dialyzer reuse, handling of medication, dialysis circuit priming, findings of water and dialysate testing, environmental reservoirs such as wall boxes, vascular access care practices, pulsed-field gel electrophoresis, and whole-genome sequencing of bacterial isolates. Outcomes of interest were cases defined by a positive blood culture for any Gram-negative bacteria drawn from July 1, 2015, to November 30, 2016, from patients who had received hemodialysis at Facility A, B, or C.

The three facilities were less than 20 miles apart and had between 12 and 36 dialysis stations. All three shared corporate ownership and some products (medication, dialyzers, acid, and bicarbonate) came from the same distributors. The same municipal water supply was used by Facilities A and B; Facility C had a different supply. The three facilities did not commonly share staff.

During the study period, there were 58 cases of bloodstream infections associated with Gram-negative organisms; the infections occurred in 51 patients. Facilities A and B had the majority of cases (n=52; 90%). In comparison, during the previous year, the three facilities identified 12 Gram-negative bloodstream infections. The most commonly identified Gram-negative organisms were S marcescens (n=21; 36%), P aeruginosa (n=12; 21%), and E cloacae (n=11; 19%).

Multiple Gram-negative organisms were isolated in 16 cases (28%). Hospital admission was required in 48 cases (83%), with median length of stay of 8 days. Fifty of the cases (86%) had central venous catheter for dialysis access. There were no associations between infections and individual staff members across or within facilities and no association between a single dialysis machine and a majority of the infections.

Case patients and the matched control-patients were similar in age, sex, and comorbid conditions. The examination of patient-specific risk factors  found a significant association between longer dialysis vintage and lower odds of infection (matched odds ratio [mOR], 0.19; 95% confidence interval [CI], 0.05-0.57). In session-specific analyses, there was a significant association between using a CVC for dialysis access and increased odds of bloodstream infection (mOR, 54.32; lower bound of the 95% CI, 12.19). Receiving dialysis after the first shift and having more than three staff members involved in care were more common among case-patients than control-patients (mOR, 2.83; 95% CI, 1.07-8.78 and mOR, 3.75; 95% CI, 1.20-15.52, respectively).

The investigation identified deficiencies in infection control at all three facilities, including inadequate aseptic technique during CVC care and multiple missed hand hygiene opportunities such as not changing gloves or not performing hand hygiene when changing gloves. There were multiple lapses in the more than 20 separate machine and station cleaning and disinfection processes, including not applying disinfectant to all surfaces or applying inadequate amount of disinfectant.

At Facility A, staff reported problems with wall boxes that emerged in early 2016 and peaked in summer 2016, including clogging and regurgitation of fluid from the drain, odors, and insect infestation. Staff were observed touching wall boxes and then going directly to CVC or other patient care without performing hand hygiene. At all three facilities, wall box basins were damp and often had visible pools of fluid, foaming, and waste fluid backing out of the drain.

Forty-three environmental samples from the three facilities underwent testing. Multiple environmental sources were found to have Gram-negative bacteria, including tap water, sinks, and surfaces. All wall boxes contained at least one of the three most common pathogens. S marcescens isolated from wall boxes were closely related to pulsed-field gel electrophoresis and whole-genome sequencing.

The investigators recommended addressing the observed lapses in infection control, including improving aseptic technique during AVC access, care, and maintenance; machine and station cleaning and disinfection, and hand hygiene.

There were some limitations to the investigation cited by the authors, including the limited number of patient isolates available for testing, the inability to sample every wall box or water source, and conducting the on-site testing after the peak of the infections.

“Providers should be aware that wall boxes are a potential source of Gram-negative bloodstream infections in dialysis settings,” the researchers said.

“Infections with Gram-negative organisms commonly found in water-related biofilms should prompt investigation into water and sources of waste fluid serving as potential reservoirs in the healthcare environment. Infection prevention and control practices should be regularly assessed and incorporated into routine quality improvement activities in all healthcare settings to decrease the likelihood of pathogen transmission from the environment to patients.”

Takeaway Points

  1. Reports to the CKC in 2016 resulted in an investigation of an outbreak of Gram-negative bloodstream infections at three outpatient dialysis facilities; the investigations focused on patients who received dialysis at one of the three facilities from July 2015 to November 2016.
  2. During the study period, 58 cases of Gram-negative bloodstream infections occurred in 51 patients; of those, 48 patients required hospitalization.
  3. The outbreak was attributed to wall boxes, a previously unidentified source of contaminated fluid and biofilms in the area providing immediate patient care.