Over the past 2 years, several reports have detailed cases of sterile site infection caused by Mycobacterium chimaera among postoperative cardiac surgery patients.Reference Sax, Bloemberg and Hasse 1 – Reference Sommerstein, Schreiber and Diekema 3 Such infections have been associated with contaminated heater-cooler units (HCUs) used in operating rooms. Transmission of infection is thought to occur via aerosolization of contaminated water from these devices.Reference Garvey, Ashford, Bradley, Martin, Walker and Jumaa 4 – Reference Chand, Lamagni and Kranzer 6
Mycobacterium chimaera is a slow-growing nontuberculous mycobacterium (NTM) belonging to the M. avium complex (MAC) (M. avium and M. intracellulare).Reference Sax, Bloemberg and Hasse 1 , Reference Sommerstein, Ruegg, Kohler, Bloemberg, Kuster and Sax 2 Mycobacerium chimaera may cause pulmonary infection, especially in patients with underlying respiratory disease, as well as skin and bone infection, and disseminated infection in the immunocompromised.Reference Garvey, Ashford, Bradley, Martin, Walker and Jumaa 4 , Reference Garvey, Bradley and Walker 5 As an environmental organism, M. chimaera has been identified in biofilms and water sources, including household water.Reference Garvey, Ashford, Bradley, Martin, Walker and Jumaa 4 , Reference Garvey, Bradley and Walker 5 Nontuberculous mycobacteria are also recognized biofilm-forming constituents of hospital water systems that have the potential to infect patients, primarily via an aerosol route.Reference Sax, Bloemberg and Hasse 1 – Reference Chand, Lamagni and Kranzer 6
Public Health England (PHE) guidelines published last month highlight the use of extracorporeal membrane oxygenation (ECMO) devices as a risk factor for M. chimaera infection. 7 It has been hypothesized that ECMO treatment might constitute a risk for transmission of water-borne pathogens,Reference Trudzinski, Schlotthauer and Kamp 8 and ECMO is a life-saving technology, particularly for patients with severe respiratory failure.Reference Trudzinski, Schlotthauer and Kamp 8 Additionally, venoarterial circuit configurations offer prolonged circulatory support, which is comparable to the short-term support provided by cardiopulmonary bypass during cardiac surgery.Reference Trudzinski, Schlotthauer and Kamp 8 While patients treated with ECMO for respiratory failure have smaller potential entry sites for pathogens than those undergoing open-chest heart surgery, they are highly immunocompromised and susceptible to opportunistic infections.Reference Trudzinski, Schlotthauer and Kamp 8 Patients may be subjected to ECMO treatment for a prolonged duration of up to several months; thus, their potential exposure to waterborne pathogens must be assessed.Reference Trudzinski, Schlotthauer and Kamp 8
University Hospitals Birmingham (UHB) NHS Foundation Trust is a tertiary referral teaching hospital in Birmingham, United Kingdom, that provides clinical services to nearly 1 million patients every year. At UHB, ECMO devices are used regularly for patients in critical care with cardiac and/or respiratory failure. Here, we report the results of weekly microbiological water sampling after decontamination following the manufacturer’s (ie, Maquet’s) guidance.
METHODS
Water Sampling
Water samples (100 mL) were collected from the water tank in the 2 Heater Unit 35 (HU35) ECMO devices (Maquet. Rastartt, Germany) used at UHB as described previously by Garvey et alReference Garvey, Ashford, Bradley, Martin, Walker and Jumaa 4 (2016). Briefly, water samples were taken weekly before and after the disinfection of each HU35, while the device was running. The neutralizer sodium thiosulphate added to all water samples, which were processed on the day of collection. Every 2 weeks, selected water samples (100 mL) were also sent on the day of testing (at 4°C) to the Mycobacterium Reference Laboratory (MRL, Public Health England, Heart of England NHS Foundation Trust, Birmingham, UK) for mycobacterium testing. These samples were assessed using standard operating procedures. 9 No enumeration of M. chimaera from water sampling was undertaken; the MRL just confirmed presence or absence of M. chimaera.
Isolation and Characterization of Gram-Negative Bacteria
Any suspected Gram-negative isolates growing on the membrane filtration agar plates were identified using matrix assisted laser desorption/ionization time-of-flight mass spectroscopy (MALDI-TOF MS, Vitek MS, bioMèrieux, Marcy-l'Étoile, France) and Vitek (Biomerieux).
The water circuit of each HU35 comprises a pump, water tank, couplings, and tubing for the HU35. Prior to 2017, tubing for each HU35 was replaced every 6 months. Since January 2017, in addition to the tubing being replaced at 6-month intervals, the tubing was replaced after every high-level disinfection. 10
Decontamination
At UHB, decontamination of each HU35 was carried out according to the manufacturer’s guidance as follows. 10
Decontamination Prior to 2017
From the summer of 2015 to December 2017, the decontamination of the ECMO unit involved draining each HU35, refilling it with filtered tap water (0.2 µm filter, Pall, New York, NY; changed every month as per manufacturer’s instructions) and addition of a Disifin tablet (buffered sodium p-toluenesulfonchloramide, Disifin, Bedfordshire, UK) to the device. The HU35 was switched on, and the tablet was circulated in the unit for 30 minutes, after which the water was removed and tank was rinsed.
Decontamination After 2017
A change in the manufacturer’s guidance in September 2016 resulted in the use of a chlorine-releasing agent. 10 At UHB, we have carried out the new guidance from January 2017 to the present. A high-level disinfection at UHB was performed on the HU35 to reduce the bioburden. This procedure was performed once because atypical mycobacteria were present in the water system. Briefly, the HU35 was emptied while the device was off, then the entire water volume of the HU 35 was refilled with a 5% chloramine-T (Chlorina, Germany) solution. According to the manufacturer’s instructions, each HU35 was disinfected for 110 minutes while the device was on. Afterward, the solution was left in the tank for 24 hours (with the unit was switched off). The HU35 was then switched on for 5 minutes, the solution was removed, and the tank was rinsed. The HU35 was running for 3 minutes during the rinse process. 10 Following this high-level disinfection, a weekly disinfection protocol using 2% chloramine-T was carried out every 7 days and after each use of the device. 10 Briefly, this procedure involved emptying the HU35 while it was switched off, disinfecting the device for 110 minutes with 2% chloramine-T while it was running, and finally rinsing the HU35 with filtered water while it was running for 3 minutes. 10
RESULTS
Water Results
Water counts from the HU35s using the pre-2017 disinfection method yielded TVCs >300 CFU per 100 mL, including a range of environmental Gram-negative isolate (e.g., Pseudomonas aeruginosa) and as non-Aspergillus molds (Table 1). Mycobacterium chimaera was cultured from water samples sent to the Mycobacterium Reference Laboratory (Table 1). With TVCs >300 CFU per 100 mL using the pre-2017 disinfection regimen, a high-level disinfection was undertaken, this time with a chlorine-releasing agent (5% chloramine-T) to reduce the bioburden in the HU35s. After the high-level disinfection procedure, water samples yielded 0 CFU per 100 mL. Ever since, weekly water samples have been taken before and after the weekly disinfection and/or disinfection after use; thus far, these weekly water samples have yielded 0 CFU per 100 mL and no atypical mycobacteria.
TABLE 1 Decontamination Results of the Extracorporeal Membrane Oxygenator Heater Unit 35 (ECMO HU35) Devices Used at UHB
NOTE: CFU, colony-forming units; +, organism(s) present; −, no organisms present.
a Weekly water samples are taken before and after the weekly disinfection or disinfection after use. Thus far, weekly water samples have yielded 0 CFU per 100 mL and no atypical mycobacteria.
DISCUSSION
Invasive M. chimaera infections associated with HCUs have been reported throughout Europe.Reference Chand, Lamagni and Kranzer 6 The risk for transmission of waterborne pathogens and subsequent infection risk posed by ECMO devices at present remains unclear. Trudzinski et alReference Trudzinski, Schlotthauer and Kamp 8 (2016) detected M. chimaera in 9 of 18 water samples from 10 different thermoregulatory ECMO devices; no mycobacteria were found in the 9 room air samples and other environmental samples. Among 118 ECMO patients, 76 had bronchial specimens analyzed for mycobacteria, and M. chimaera was found in 3 individuals without signs of mycobacterial infection at the time of sampling.Reference Trudzinski, Schlotthauer and Kamp 8 Trudzinski et alReference Trudzinski, Schlotthauer and Kamp 8 (2016) concluded that M. chimaera can be detected in water samples from ECMO-associated thermoregulatory devices and might potentially put patients at risk of infection.
In patients presenting with severe acute respiratory distress syndrome (ARDS) who become unresponsive to conventional treatment, the use of ECMO, which temporary replaces pulmonary function, can represent a life-saving alternative while the underlying cause is resolved.Reference Trudzinski, Schlotthauer and Kamp 8 An ECMO can be used as a rescue therapy to avoid injurious effects of mechanical ventilation and to rescue patients from extreme gasometrical alterations.Reference Trudzinski, Schlotthauer and Kamp 8 Data from our study show that the ECMO devices used at UHB were contaminated with environmental Gram-negative bacteria (eg, Pseudomonas species, atypical mycobacteria, and fungi), which may pose an infection risk in cardiac patients. Atypical mycobacteria can be detected in household water and water-containing medical devices,Reference Garvey, Ashford, Bradley, Martin, Walker and Jumaa 4 , Reference Trudzinski, Schlotthauer and Kamp 8 and it has been shown that the LivaNova 3T HCUs, which use water for thermoregulation during cardioplegia, are responsible for the recent M. chimaera outbreak worldwide.Reference Sax, Bloemberg and Hasse 1 , Reference Chand, Lamagni and Kranzer 6 Chand et alReference Chand, Lamagni and Kranzer 6 (2017) postulated that a point source for the contamination of HCUs had occurred, that is, that the LivaNova 3T units were contaminated within the manufacturing site.Reference Chand, Lamagni and Kranzer 6 Contamination of the ECMOs could have occurred at the site of manufacturing, like the LivaNova 3T units, or from water supplies at our hospital.Reference Sax, Bloemberg and Hasse 1 , Reference Chand, Lamagni and Kranzer 6 Further work on M. chimaera isolates related to ECMO devices is required. Here, we show that, similarly to HCUs, decontamination procedures are essential to reduce the microbial load in the ECMO devices and that weekly microbiological sampling is required to monitor the bioburden in the ECMO devices. On initial examination, the ECMO devices had a high microbial load, with TVC >300 CFU per 100 mL. The TVC results following disinfection using the pre-2017 protocol showed that this method was inadequate. The disinfection regimen implemented in 2017 has been effective in reducing the bioburden. However, the ongoing maintenance to eliminate biofilm requires additional resources. Clearly, the increased disinfection and sampling procedures recommended by the manufacturer have added a considerable burden of duties and costs for perfusionists, cardiology, and surgical staff. The decontamination guidance recommended by the manufacturers is time-consuming, taking up to 24 hours for a high-level disinfection. Garvey et alReference Garvey, Ashford, Bradley, Martin, Walker and Jumaa 4 , Reference Garvey, Bradley and Walker 5 (2016, 2017) have shown that enhanced decontamination of HCUs has led to degradation of the internal cooper fixings, and a similar affect could be seen with the ECMO devices. Previous studies have shown that HCUs aerosolize microorganisms present in the HCUs themselves such as M. chimaera.Reference Sommerstein, Ruegg, Kohler, Bloemberg, Kuster and Sax 2 , Reference Sommerstein, Schreiber and Diekema 3 It has been postulated that ECMO devices might be associated with the similar problem.Reference Trudzinski, Schlotthauer and Kamp 8 In this study, we did not attempt to demonstrate that the organisms detected in the HU35s had aerosolized. Trudizinski et alReference Trudzinski, Schlotthauer and Kamp 8 (2016) undertook environmental sampling and did not detect any aerosolization. At present, no conclusive link of transmission from the ECMO units to patients has been confirmed. Further investigation of ECMO devices is needed to better understand whether transmission occurs via these devices. To our knowledge, this is the first report to focus on the decontamination of ECMO devices.
The risk of infections from M. chimaera is still low in the context of other risks; transmission from an ECMO device to a patient has not yet been described.Reference Chand, Lamagni and Kranzer 6 However, a heavily biofouled ECMO may have the potential to aerosolize like the HCUs.Reference Sax, Bloemberg and Hasse 1 – Reference Sommerstein, Schreiber and Diekema 3 The decontamination of these devices is likely to have an essential role in mitigating risk, and this report highlights the fact that technique is important. Clinicians need to consider the mitigation of infection risk versus the need to maintain a patient’s respiratory and cardiac function using ECMO devices.Reference Chand, Lamagni and Kranzer 6
ACKNOWLEDGMENTS
We thank the Infection Prevention and Control and Clinical Perfusion Teams at the University Hospitals Birmingham NHS Foundation Trust.
Financial support: No financial support was provided relevant to this article.
Potential conflicts of interest: All authors report no conflicts of interest relevant to this article.
