Mycobacterium chimaera infections in post–operative patients


Volume 43-5, May 4, 2017: Implementation science


Mycobacterium chimaera infections in post–operative patients exposed to heater–cooler devices: An overview

T Ogunremi1*, G Taylor2, L Johnston3, K Amaratunga1,4, M Muller5, A Coady1, K Defalco1, K Dunn1, J Johnstone6, S Smith2, J Embree7, B Henry8, J Stafford9 on behalf of the Infection Prevention and Control Expert Working Group10


1 Centre for Communicable Diseases and Infection Control, Public Health Agency of Canada, Ottawa, ON

2 Department of Medicine, University of Alberta, Edmonton, AB

3 Queen Elizabeth II Health Science Centre and Dalhousie University, Halifax, NS

4 Department of Medicine, University of Ottawa, Ottawa, ON

5 St. Michael’s Hospital, Toronto, ON

6 Public Health Ontario, Toronto, ON

7 University of Manitoba, Winnipeg, MB

8 Office of the Provincial Health Officer, Ministry of Health, Victoria, BC

9 Department of Health, Government of New Brunswick, Fredericton, NB

10 All working group contributors are noted at the end of the paper


Suggested citation

Ogunremi T, Taylor G, Johnston L, Amaratunga K, Muller M, Coady A, Defalco K, Dunn K, Johnstone J, Smith S, Embree J, Henry B, Stafford J. Mycobacterium chimaera infections in post-operative patients exposed to heater–cooler devices: An overview. Can Commun Dis Rep. 2017;43(5):107-13.


A multi-country outbreak of Mycobacterium chimaera infection associated with contaminated heater–cooler devices (HCDs) has been reported, with more than 70 cases in Europe and the United States and two cases in Canada to date. The epidemiological and microbiological characteristics of this outbreak provide evidence for common-source transmission of M. chimaera from the exhaust air of intrinsically contaminated HCDs to patients during cardiac surgery. To date, all reported cases have been associated with Stöckert 3T HCDs manufactured at one plant by LivaNova prior to September 2014. Implantation of prosthetic material increases the risk of infection. Infections usually present as prosthetic valve endocarditis, vascular graft infection or disseminated infection. Reported mortality rates have varied, but were often over 40%.

Several measures are recommended to facilitate case-finding and mitigate risk of exposure. The feasibility of some risk mitigation measures and their effectiveness in reducing the risk of exposure are yet to be determined. Until HCDs are redesigned in a manner that prevents water contamination and aerosolization, separating the HCD exhaust air from the operating room air during surgery may be the most effective risk mitigation strategy. However, possible unintended consequences of this approach should be considered. This overview summarizes findings from peer-reviewed and other relevant national documents on key features of the outbreak, including the source, identified risk factors for infection, signs and symptoms of infection, burden of disease, risk mitigation measures, management challenges and knowledge gaps.

Keywords: Mycobacterium chimaera, heater–cooler device, cardiac surgery, cardiopulmonary bypass


Health care–associated infections related to medical device contamination and biofilm formation have been documented in the literature Footnote 1. Recently, heater–cooler devices (HCDs) used during cardiopulmonary bypass (CPB) for cardiac surgeries and during extracorporeal membrane oxygenation (ECMO) have come under scrutiny due to infections linked to contaminated devices Footnote 2Footnote 3.

Heater-cooler devices have water tanks that pump temperature-controlled water through closed circuits to external heat exchangers that regulate patient body temperature by convection Footnote 4. The device is equipped with a radiator and fan to facilitate constant dissipation of excess heat through grid openings and the stirring of water in the tank results in aerosolization via the exhaust air Footnote 4Footnote 5. HCDs are subject to biofilm formation. A biofilm is an aggregate of microorganisms embedded within an extracellular matrix that adhere to each other and to internal surfaces, such as the interior of HCDs.

Several types of microorganisms have been isolated from contaminated HCDs, including nontuberculous mycobacteria (NTM), which are ubiquitous in soil and water and have been linked to health care-associated infections Footnote 6Footnote 7Footnote 8Footnote 9. Investigations of NTM infection clusters following cardiac surgery detected Mycobacterium chimaera as the causative microorganism. M. chimaera is a slow-growing NTM included in the Mycobacterium avium complex Footnote 3Footnote 9Footnote 10Footnote 11Footnote 12. It is less susceptible to disinfection procedures due to its cell wall constituents and its ability to form biofilms. Isolation and identification of M. chimaera from clinical specimens requires specialized microbiological techniques Footnote 3. Transmission was associated with a single model of HCD manufactured by Sorin (now LivaNova) Footnote 3Footnote 13. Cultures from HCD water tanks, water circuits and air samples taken while HCDs were in use have grown M. chimaera Footnote 5Footnote 8Footnote 9Footnote 11.

Although M. chimaera contamination of ECMO devices has been reported, contamination did not spread to the air in the room while the devices were running and no ECMO-associated M. chimaera infections were reported Footnote 2Footnote 14. Nonetheless, the need to assess potential patient exposure from ECMO has been recognized since patients treated with ECMO, who are often critically ill and highly immunocompromised, may be exposed to the device for an extended period of time Footnote 2. National guidance documents and safety communications describing risk mitigation measures and testing recommendations in Canada, United Kingdom (UK), United States (US) and Australia have been published Footnote 13Footnote 15Footnote 16Footnote 17Footnote 18Footnote 19Footnote 20.

The objective of this overview is to summarize relevant literature on the current multi-country outbreak of M. chimaera infection. The source of exposure, risk factors for infection, signs and symptoms of infection, disease burden, risk mitigation measures, challenges and gaps are summarized. This overview may be a helpful resource for Canadian health care facilities and providers who use HCDs. It may also support informed decision-making by authorities responsible for implementing infection prevention and control measures.


A worldwide literature search was undertaken by the Health Library (Health Canada) using Ovid MEDLINE, EMBASE and Global Health databases for studies published from January 1, 2007 to March 8, 2017. The search strategy was developed using database-specific thesauri for “Mycobacterium chimaera”, “heater-cooler devices”, and “cardiac surgery”. The search was limited to studies in English and French with no filters applied to limit retrieval by study design. A grey literature search was also conducted by the Health Library to identify relevant national guidance documents and safety communications. The reference lists of relevant guidance documents were hand searched for additional relevant studies.

Full texts of all relevant studies were screened to identify those reporting on HCD-associated M. chimaera infection in postoperative cardiac surgery patients and any risk mitigation measures described. A narrative synthesis of the relevant peer-reviewed publications, national guidance documents and/or safety communications was done.


A total of 95 articles were retrieved from peer-reviewed and grey literature searches, including a reference list search of identified documents. Information from 38 relevant documents was included in this overview. Fifty-seven articles were excluded for one of several reasons including studies that reported on case(s) already described in detail elsewhere; studies that focused on NTM in general (not specifically M. chimaera); studies that did not discuss patient exposure or transmission; and national guidance documents or safety communications that did not provide additional information to that obtained from similar documents from Canada, the US, Australia and Europe.

Source of exposure

To date, all cases of M. chimaera infection reported internationally have been associated with Stöckert 3T HCDs manufactured in Germany by LivaNova before September 2014 Footnote 3Footnote 9Footnote 13Footnote 15Footnote 21Footnote 22Footnote 23. Phylogenetic analysis by whole genome sequencing and other means showed that isolates from infected patients and from water and exhaust air of used and new Stöckert 3T HCDs were closely related, suggesting global distribution of contaminated HCDs and a hospital-independent, common source for the current outbreak Footnote 5Footnote 9Footnote 12Footnote 22Footnote 24Footnote 25Footnote 26. LivaNova implemented changes to their disinfection processes in an attempt to reduce the risk of M. chimaera contamination of 3T HCDs manufactured after September 2014 Footnote 13Footnote 15Footnote 27Footnote 28. Tests conducted on HCDs manufactured by a different company detected M. chimaera in the water but not in air samples, and the isolate obtained was genetically distinct from isolates obtained from Stöckert 3T HCDs Footnote 12Footnote 25Footnote 29.

During surgeries, the HCD is often positioned adjacent to the cardiopulmonary bypass machine and the patient. Recently, one of the considerations related to minimizing patient exposure to exhaust air from the HCD has to do with the feasibility of positioning the HCD immediately beside the floor-level exhaust in the operating room.

Risk factors for infection

Cases of M. chimaera infection following exposure to HCDs during cardiopulmonary bypass have been reported in patients who had undergone surgery in Europe (UK, France, Switzerland, Netherlands, Germany, Ireland and Spain) as well as in the US, Australia, Canada and Hong Kong Special Administrative Region Footnote 18. Patients undergoing cardiac surgery involving cardiopulmonary bypass where body temperature is regulated by HCDs are at risk of exposure and infection Footnote 8. Patients undergoing cardiopulmonary bypass for over two hours had higher odds of NTM infection (odds ratio: 16.5; 95% CI: 3.2–84) Footnote 8. In hospitals where at least one HCD-associated M. chimaera infection was identified, the risk of a patient getting an infection was approximately 0.1–1% Footnote 11Footnote 30Footnote 31. Of 115,664 surgical procedures in England involving repair or replacement of cardiac valves (between 2007 and 2014), the risk of NTM infections increased from less than 0.2/10,000 person-years before 2010 to 1.65/10,000 person-years in 2013 Footnote 29.

Implantation of prosthetic material (e.g., heart valve, vascular graft, left ventricular assist device) increased the risk of infection Footnote 3Footnote 11Footnote 13Footnote 29. Limited data suggest that heart transplants may also increase the risk of infection Footnote 3Footnote 32.

No case has occurred in operating room personnel exposed to aerosolization from HCDs.

Signs and symptoms of infection

Signs and/or symptoms of invasive M. chimaera infection following exposure to aerosols from an HCD may not occur for months or years after exposure, with a mean time between exposure and diagnosis of 1.6 years (range: 0.1–6.3 years) Footnote 3Footnote 10Footnote 14Footnote 23Footnote 32. The infection usually presents as prosthetic valve endocarditis, vascular graft infection or disseminated infection although a variety of extracardiac sites may also be infected (Table 1Footnote 9Footnote 10Footnote 11Footnote 13Footnote 18Footnote 29Footnote 33. Clinical manifestations of infection are diverse and symptoms may be nonspecific Footnote 12Footnote 23. In some cases, extracardiac manifestations preceded cardiovascular disease Footnote 11. A description of a compatible syndrome for NTM infection published by the Canadian Public Health Laboratory Network (CPHLN) is shown in Table 1 Footnote 16.

Table 1: Clinical symptoms of patients with Mycobacterium chimaera infection
Type of symptoms Clinical symptoms
Constitutional Recurrent or prolonged fever, fatigue, shortness of breath, weight loss, night sweats
Cardiac Prosthetic valve endocarditis and/or prosthetic vascular graft infection
Extracardiac Bone infection, sternotomy surgical wound infection, mediastinitis, hepatitis, bloodstream infection, ocular infection (panuveitis, multifocal choroiditis, chorioretinitis)
Immunologic/embolic Splenomegaly, cytopenia
Infants Febrile episodes and failure to thrive

Disease burden

M. chimaera infection requires aggressive medical treatment with combination antimycobacterial therapy and sometimes repeat surgical intervention. The infection generally results in substantial morbidity with prolonged hospitalization, adverse effects of medical and surgical treatment, and/or treatment failure Footnote 3Footnote 11Footnote 18Footnote 29. In Europe, at least 52 cases have been reported as of January 2017 Footnote 12Footnote 18. Three cases have been identified in Australia, 24 in the US and two in Canada Footnote 20Footnote 23Footnote 32Footnote 34. Individual patient information was not always reported. From the data available, most cases were in older adults although patient age ranged from one to 81 years old, including two pediatric patients. Approximately 83% of the patients were male. Most studies reported a mortality rate over 40% (see Table 2Footnote 3Footnote 11Footnote 12Footnote 29Footnote 32 and mortality was high when significant delays in diagnosis occurred and patients were severely ill when appropriate antimycobacterial treatment was implemented. It remains unclear whether increased awareness and earlier diagnosis will reduce the mortality associated with M. chimaera infection.

Table 2: Reported mortality from Mycobacterium chimaera infection associated with heater–cooler devices
Reference (country / region) Number of patients diagnosed Number of deaths (mortality) %
Kohler et al., 2015 (Europe) Footnote 11 10Table 2 - Footnote 1 4 (40%)Table 2 - Footnote 2
Chand et al., 2016 (Europe) Footnote 29 18Table 2 - Footnote 3 9 (50%)
Appenheimer et al., 2016 (US) Footnote 32 24 NR (46%)Table 2 - Footnote 4
European Centre for Disease Prevention and Control, 2016 (Europe) Footnote 18 52Table 2 - Footnote 5 10 (<19%)Table 2 - Footnote 6
Haller et al., 2016 (Germany) Footnote 9 5 1 (20%)Table 2 - Footnote 7
Tan et al., 2016 (US) Footnote 33 3 2 (67%)Table 2 - Footnote 6,Table 2 - Footnote 7
Public Health England, 2017 (Europe) Footnote 12 26 15 (58%)
Australian Commission on Safety and Quality in Health Care, 2017 (Australia) Footnote 20 3 0 (0%)

Risk mitigation measures

Key measures identified to facilitate case-finding and mitigate future exposure to M. chimaera are summarized in Table 3.

Table 3: Recommended measures to facilitate case-finding and mitigate future risk of Mycobacterium chimaera exposure
Risk mitigation measure Additional context and/or limitation
Health care provider notification and education Footnote 11Footnote 12Footnote 28Footnote 32
  • Cases have been detected via provider notification.
  • Earliest implicated surgery was performed in 2007.
  • Maintain high clinical suspicion for M. chimaera or other NTM infection in patients (who underwent surgery involving CPB with use of HCDs from 2007 to implementation of risk mitigation measures).
Patient notification Footnote 8Footnote 12Footnote 28Footnote 32
  • To date, no cases have been identified via patient notification.
  • Testing is not recommended for asymptomatic exposed individuals.
  • Until effective risk mitigation measures are implemented, information regarding potential exposure should be provided to patients prior to surgery.
Enhanced prospective NTM surveillance Footnote 9Footnote 21
  • The ECDC has published a protocol for case detection.
Ensure traceability of HCDs in use Footnote 12
  • Individual units used in each surgery should be recorded in the event of a later infection.
Remove potentially contaminated HCDs from service Footnote 12Footnote 15Footnote 27
  • Where possible, all Stöckert 3T HCDs manufactured by LivaNova prior to September 2014 should be removed from service.
  • In some settings, risk of deferring surgery exceeds risk of surgery with use of proven or suspect contaminated HCD.
Replace contaminated HCDs, plus accessories, tubing and connectors, to prevent recontamination Footnote 13Footnote 15Footnote 27Footnote 35
  • LivaNova implemented a program to, in some circumstances, provide users with a loaner device to continue surgical procedures while their devices are undergoing deep disinfection. International demands for replacement of HCDs may result in a backlog in supply.
Use manufacturer’s operation protocol including updated cleaning and disinfection procedures Footnote 3Footnote 9Footnote 12Footnote 15Footnote 27Footnote 28Footnote 35
  • Maintain log of cleaning and disinfection of HCDs.
  • Regularly check manufacturer’s website for relevant updates.
  • Current decontamination protocols are yet to be validated. Studies have challenged the effectiveness of these protocols, suggesting a systematic decontamination failure. Biofilm removal is essential for effective decontamination of HCDs.
Routine microbiological testing of HCDs in use Footnote 12Footnote 15Footnote 17Footnote 25Footnote 27Footnote 36
  • This is not widely adopted because of the high rate of false negative results and the lack of standardized and validated methods for sample collection, processing and detection of M. chimaera.
  • The Canadian Public Health Laboratory Network and the US FDA advise against obtaining routine environmental cultures from HCDs for M. chimaera.
Apply engineering solutions to enable reliable separation of HCD exhaust air from operating room air Footnote 4Footnote 5Footnote 12Footnote 13Footnote 15Footnote 18Footnote 25Footnote 26Footnote 37 Options include:
  • Place the HCD outside the operating room with tubing connected through an opening in the wall (ensuring operating room positive air pressure is maintained). Although this is the most reliable solution, the unintended consequences of this solution (e.g., possibly altered airflow in operating rooms and a longer distance between the HCD and surgical field) are unknown.
  • Encase the HCD in custom-made housing with separate ventilation (e.g., connected to the operating room exhaust conduit). Attachments to the HCD may need to be approved by the manufacturer. The unintended consequences of this solution (e.g., effects of custom-made housing on how well the device functions) are unknown.
  • If unable to reliably separate HCD exhaust air from operating room air, move the HCD as far as possible (preferably more than five metres) from the surgical field with the vent exhaust directed away from both the patient and the exposed instruments, and if possible, place the HCD close to the room air exhaust. Smoke dispersal experiments demonstrated that exhaust air from HCDs was propelled to merge with ultraclean airflow near the ceiling of the operating room. As a result, it is unclear whether this approach is useful in separating HCD exhaust air from operating room air Footnote 4.

Challenges and gaps

Table 4 summarizes the challenges and gaps in evidence informing the clinical management of M. chimaera infection.

Table 4: Challenges and gaps in evidence informing the management of Mycobacterium chimaera infection
Challenge / gap Additional context
The magnitude of the risk of M. chimaera infection and the extent of the outbreak is unknown Footnote 12Footnote 14Footnote 29
  • High prevalence of M. chimaera in HCDs has been reported (up to 80% in Denmark).
  • The risk of patient infection currently appears to be low; however, if infection occurs, the impact on the patient could be severe.
  • The risk of infection and clinical presentation among the pediatric patient population is unknown.
Delay in symptom onset and diagnosis of infection Footnote 3Footnote 10Footnote 14Footnote 23Footnote 32
  • Documented time from exposure to diagnosis was between 0.1–6.3 years (mean of 1.6 years).
  • Few laboratories are equipped to culture and identify M. chimaera, which could contribute to delay in diagnosis.
  • Slow growth of M. chimaera culture contributes to delayed diagnosis.
  • Early collection of dedicated mycobacterial culture can result in diagnosis within a shorter timeframe than is commonly reported.
Effectiveness and adverse effects of therapy Footnote 3Footnote 11Footnote 29Footnote 32
  • M. chimaera infection can be very difficult to treat due to the microorganism’s intrinsic resistance to many antimicrobial agents; its propensity for biofilm formation on implanted devices; and deep-seated infection sites that are challenging for antimicrobial penetration (e.g., endocarditis, graft infection and bone).
  • Therapy is prolonged and requires a combination of antimicrobial agents.
  • Disseminated infection has often required repeat surgical interventions with high mortality rates reported.
Development of new HCD designs is pending Footnote 5Footnote 15
  • Construction of custom-built containers with HEPA filters to house HCDs that cannot be placed outside the operating room is underway, but their effectiveness is currently unknown.
  • HCD manufacturers are modifying HCD designs to limit aerosolization and prevent transmission.
Extent of HCD-associated infections caused by other microorganisms such as Legionella species is unknown Footnote 12Footnote 29
  • National surveillance in the UK (2007–2016) did not identify any cases of Legionnaire’s disease in health care workers with potential occupational exposure to HCDs.
  • Postoperative cardiac surgery endocarditis due to Legionella species has not been reported during this outbreak.


Findings from this overview indicate a low but increased risk of M. chimaera infection with use of common source-contaminated HCDs during CPB Footnote 29. Given the long latency period, additional cases are expected. The true magnitude of risk following exposure is uncertain; current estimates are based on very limited data. Nonetheless, the risk of delaying cardiac surgery is generally considered far greater than the risk posed by this infection, even when the infection risk has not been entirely mitigated Footnote 28.

Future patient exposure may be prevented by implementing risk mitigation measures, including the use of uncontaminated HCDs or replacement of contaminated HCDs as soon as possible. Case finding may be expedited by the development of polymerase chain reaction (PCR)-based assays for rapid and reliable detection of M. chimaera in clinical or environmental samples.

Improved HCD designs that facilitate reliable decontamination and prevent aerosols from reaching the operative field are urgently needed Footnote 5Footnote 11. These developments may require collaborative discussions between medical device manufacturers, engineers and infection prevention and control experts.

This overview is limited by insufficient data to estimate the true magnitude of risk of infection and absence of data on efficacy and feasibility of risk mitigation measures.


The epidemiological and microbiological characteristics of this outbreak provide evidence for transmission of M. chimaera from the exhaust air of contaminated Stöckert 3T HCDs to patients during CPB, resulting in endocarditis, surgical site infections and/or disseminated infections. The true magnitude of risk following exposure is uncertain; it is currently estimated based on very limited data.

Strategies that separate the HCD exhaust air from the operating room air during surgery may be the most effective risk mitigation measures. The feasibility of implementing currently recommended risk mitigation measures is yet to be determined and studies are needed to determine if there are any unintended consequences from implementation of these measures. Development of HCDs with new designs that are airtight and/or not susceptible to biofilm formation may help address this problem.

Authors’ statement

TO – project administration, conceptualization, methodology, research, data abstracting and writing (original draft, review and editing); GT, LJ, MM – clinical expertise, intellectual content, review, editing and writing; KA - clinical expertise, intellectual content, review and editing; AC, K Defalco – review of abstracted data, research, review and editing; K Dunn – conceptualization, supervision, review and editing; JJ, SS, JE, BH, JS – clinical expertise, intellectual content, review and editing.

Conflict of interest



The authors would like to thank the following members of the Infection Prevention and Control Expert Working Group for their contributions to discussions on the content of this manuscript:

  • Gwen Cerkowniak, Saskatoon Health Region, Saskatoon, SK
  • Maureen Cividino, Public Health Ontario, Toronto, ON
  • Della Gregoraschuk, Alberta Health Services, Edmonton, AB
  • Patsy Rawding, Nova Scotia Health Authority, Wolfville, NS
  • Sandra Savery, Centre de santé et services sociaux des Sommets, Sainte-Agathe-des-Monts, QC
  • Heidi Pitfield, Simcoe Muskoka District Health Unit, Barrie, ON
  • Patrice Savard, Centre hospitalier de l’Université de Montréal, Montréal, QC


The authors would like to thank the Canadian Public Health Laboratory Network (CPHLN) for providing information presented in Table 1 of this manuscript.

The authors would also like to thank Lynda Gamble (Health Library, Health Canada) for conducting the literature search for this overview; Caroline M. Desjardins (Public Health Agency of Canada) for citing and generating references for this manuscript; and Dr. Margaret Gale-Rowe (Public Health Agency of Canada) for reviewing the final version of the manuscript.


This work was supported by the Public Health Agency of Canada.

Page details

Date modified: