Patients undergoing extracorporeal membrane oxygenation (ECMO) are at increased risk of infection. We present the first known report of nosocomial infection with Cupriavidus pauculus attributable to contamination from ECMO equipment and describe the measures taken to halt subsequent infections. A cluster of infections in ECMO patients should prompt team members to consider contamination of equipment with environmental pathogens as a possible cause.
From the *University of Arkansas for Medical Sciences and †Arkansas Children's Hospital, Little Rock, Arkansas.
Submitted for consideration June 2010; accepted for publication in revised form July 2010.
Reprint Requests: Stephanie H. Stovall, MD, University of Arkansas for Medical Sciences, 1 Children's Way, Slot #512-11, Little Rock 72202, AR. Email: firstname.lastname@example.org.
Extracorporeal membrane oxygenation (ECMO) is widely used for supporting patients with refractory cardiorespiratory failure. Patients supported with ECMO have substantial risk of hospital-acquired infection.1,2 Previous studies have linked infection in ECMO patients with longer duration of support, but pathogens are similar to those found in conventional intensive care settings.3,4 A cluster of similar organisms in a select patient group, such as ECMO patients, should alert clinicians to consider common equipment or environmental sources of infection.
We describe the first reported outbreak of Cupriavidus pauculus associated with contamination of the thermoregulator reservoir in the ECMO circuit to alert other institutions to the possibility of nosocomial transmission of pathogens in this manner.
The microbiology laboratory at Arkansas Children's Hospital (ACH) isolated an unusual Gram-negative organism from multiple blood cultures submitted from one patient. Automated microbiological systems and routine biochemical testing failed to identify the organism. A microbiologist recalled previous isolation of a similar organism, identified as C. pauculus by a reference laboratory, from two other patients within the previous year. This prompted review of records and an infection control investigation to find a common source.
Cases were identified by review of microbiological records for any culture positive for C. pauculus. Medical records were reviewed in detail. Four cases of C. pauculus infection were identified, all in patients supported with ECMO.
Data were deidentified, and no communication with patients or families was attempted. The University of Arkansas for Medical Sciences Institutional Review Board classified this study as exempt.
A healthy 4-month-old infant developed fever, vomiting, and lethargy, prompting evaluation at an outside institution. He developed junctional ectopic tachycardia and refractory cardiogenic shock requiring ECMO support and was transported to ACH for evaluation for heart transplantation. Two days later, catheter cryoablation terminated his arrhythmia. Cardiac function improved, and he was weaned from ECMO after 10 days. Daily blood cultures were submitted per ECMO protocol. Blood cultures drawn on days 2 and 3 became positive for Gram-negative rods; however, no identification was determined by the BD Phoenix (Becton Dickinson; Franklin Lakes, NJ) or by laboratory biochemical tests. A reference laboratory identified the organism as C. pauculus. The patient received meropenem and gentamicin for 6 days and 3 days of piperacillin/tazobactam; all subsequent blood cultures were negative. He was discharged 14 days later.
A 3-year-old Native American boy with a history of Shone's complex was admitted to another institution with severe congestive heart failure. ECMO support was initiated and he was transported by air to ACH for transplant evaluation. Blood cultures submitted after admission grew C. pauculus. Daily blood cultures from hospital days 3–6 were also positive for C. pauculus. Treatment included meropenem until preliminary susceptibility tests showed resistance to meropenem. Definitive treatment included cefepime and piperacillin/tazobactam for 17 days. Subsequent daily cultures were negative. The patient weaned from ECMO on hospital day 8 and was discharged 2 months later after receiving a heart transplant.
A term newborn was transferred to ACH for care of hypoplastic left heart syndrome. On the fourth day after birth, a Norwood procedure with Sano modification was performed; however, the patient required postoperative ECMO support because of low cardiac output. Blood cultures drawn hours after initiation of ECMO support grew C. pauculus. Repeat blood cultures were positive for the next 2 days. The patient was treated with cefepime and ciprofloxacin for 21 days. Daily blood cultures became negative after 3 days of treatment. The patient later developed sepsis due to Candida parapsilosis and died. Autopsy revealed positive cultures for C. parapsilosis but no C. pauculus.
A 16-month-old infant with double outlet right ventricle, mitral stenosis, and pulmonary stenosis underwent surgery at another institution but was unable to wean from cardiopulmonary bypass. She was transferred to ACH while on ECMO support for heart transplantation evaluation. She developed fever on hospital day 4 and blood culture grew C. pauculus on hospital day 9. She weaned from ECMO on day 5 and was treated with meropenem and gentamicin. After C. pauculus was suspected, meropenem was changed to cefepime based on previous susceptibility patterns. Fourteen days of cefepime and 13 days of gentamicin were given with no further positive blood cultures before discharge after receiving an orthotopic heart transplant.
Investigation of the outbreak by a multidisciplinary committee determined that multiple ECMO thermoregulator reservoirs (Figure 1) were contaminated with C. pauculus. The mechanism of patient infection was suspected to be through splashing of surfaces with contaminated water during manipulation of the reservoir and transmission from the ECMO specialist's gloves to the patient's blood when using blood ports of the ECMO circuit. The thermoregulator reservoir contains water used by the circuit's heat exchanger to rewarm the patient's blood. Without this mechanism, profound hypothermia would occur. At no time the is reservoir water in contact with blood in the circuit. The reservoir is filled with water before initiation of ECMO support, and the procedure required opening of the reservoir to check its water level every shift.
A multidisciplinary team evaluated means of disinfecting the thermoregulator reservoirs that are not disposable parts of the ECMO equipment. In addition to the disinfection procedure, ECMO patient care procedures were altered to address the problem. First, ECMO specialists assembling new circuits were required to fill the thermoregulator reservoir while wearing a mask and sterile gloves, which are discarded before further manipulation of circuit components. Second, ECMO care policy now restricts opening of the reservoir during ECMO support, because the need to add water to the reservoir is extremely rare based on institutional experience. Finally, a protective cover for the reservoir was developed to minimize risk of splash contamination of the ECMO circuit in the event that water was added. During the 12 months since the last infection with C. pauculus, 51 patients have been supported by ECMO at our facility for 8,371 hours, and no further infections have occurred.
Patients undergoing ECMO support are at substantial risk of hospital-acquired infections. Higher rates of bacteremia are associated with longer duration of ECMO support.3,5 However, few reports of contamination of components of the ECMO equipment exist in the literature.
Although the thermoregulator reservoir has no connection to the blood flowing through the ECMO circuit, its proximity to sites in the circuit from which blood is accessed may provide a source of infection. A previous case report found identical isolates of Ralstonia pickettii in the thermoregulator reservoir and in a patient supported with ECMO.6 In that case, the reservoir was disinfected by a procedure recommended by the manufacturer, and no subsequent infections were reported, but the length of time after the last culture is not reported.
To our knowledge, bacteremia with C. pauculus has not been previously described during ECMO support. The multidisciplinary approach to infection control described here, including changes to patient care procedure and changes to the thermoregulator disinfection procedure, have stopped further nosocomial infection with this organism. This report underscores the need to evaluate procedures involving complex medical equipment and care processes to minimize potential nosocomial pathogen exposures in critically ill patients.
The authors thank Dr. J. LiPuma, and the Burkholderia cepacia Research Laboratory, and Repository at the University of Michigan for identifying the isolates referred for testing.
1. Burket JS, Bartlett RH, Vander Hyde K, Chenoweth CE: Nosocomial infections in adult patients undergoing extracorporeal membrane oxygenation. Clin Infect Dis
28: 828–833, 1999.
2. Odetola FO, Moler FW, Dechert RE, et al
: Nosocomial catheter-related bloodstream infections in a pediatric intensive care unit: Risk and rates associated with various intravascular technologies. Pediatr Crit Care Med
4: 432–436, 2003.
3. Schutze GE, Heulitt MJ: Infections during extracorporeal life support. J Pediatr Surg
30: 809–812, 1995.
4. Steiner CK, Stewart DL, Bond SJ, et al
: Predictors of acquiring a nosocomial bloodstream infection on extracorporeal membrane oxygenation. J Pediatr Surg
36: 487–492, 2001.
5. O'Neill JM, Schutze GE, Heulitt MJ, et al
: Nosocomial infections during extracorporeal membrane oxygenation. Intensive Care Med
27: 1247–1253, 2001.
6. Forgie S, Kirkland T, Rennie R, et al
: Ralstonia pickettii bacteremia associated with pediatric extracorporeal membrane oxygenation therapy in a Canadian hospital. Infect Control Hosp Epidemiol
28: 1016–1018, 2007.