Nosocomial Infections Due to Nontuberculous Mycobacteria
+Author Affiliations
- Reprints or correspondence: Dr. C. Fordham von Reyn, Dartmouth-Hitchcock Medical Center, Infectious Disease Section, Lebanon, NH 03756 (c.fordham.von.reyn@hitchcock.org)
Abstract
Nontuberculous mycobacteria (NTM) are ubiquitous in the environment and cause colonization, infection, and pseudo-outbreaks in health care settings. Data suggest that the frequency of nosocomial outbreaks due to NTM may be increasing, and reduced hot water temperatures may be partly responsible for this phenomenon. Attention to adequate high-level disinfection of medical devices and the use of sterile reagents and biologicals will prevent most outbreaks. Because NTM cannot be eliminated from the hospital environment, and because they present an ongoing potential for infection, NTM should be considered in all cases of nosocomial infection, and careful surveillance must be used to identify potential outbreaks. Analysis of the species of NTM and the specimen source may assist in determining the significance of a cluster of isolates. Once an outbreak or pseudo-outbreak is suspected, molecular techniques should be applied promptly to determine the source and identify appropriate control measures.
Nontuberculous mycobacteria (NTM) are found in the natural environment throughout the world and include more than 65 different species. Evidence suggests that nosocomial transmission of these organisms is increasing, which results in conditions ranging from harmless colonization to invasive infection. NTM may also contaminate microbiological specimens, which leads to unnecessary therapy and potentially harmful diagnostic procedures. Recently, investigation of putative nosocomial outbreaks of NTM has been aided by the use of molecular techniques to identify the source and mode of transmission. In the present article, we review the ecology, epidemiology, and control of nosocomial infections due to NTM.
Ecology And Microbiology
NTM exist widely in soil and water and are known to colonize and cause disease in a broad range of animal species. Environmental studies have isolated NTM from natural and potable water sources in both temperate and tropical regions [1,2,3,4,5–6]. Studies on potable or treated water indicate that some species of NTM, such as Mycobacterium xenopi and Mycobacterium kansasii, have only been isolated from man-made water systems [7]. Water obtained from municipal treatment facilities, hospitals, and homes grew NTM in 10%–95% of samples in European and American surveys. Water samples obtained from homes typically had lower rates of cultures that yielded NTM than did hospitals [8–10]. Older water systems may have higher colonization rates [11, 12], but nosocomial outbreaks related to NTM colonization of newly built water systems have been also reported [13]. Examination of a wide range of plumbing surfaces has revealed a higher density of NTM on plastic and rubber when compared with other materials, regardless of duration of exposure [14].
The highest rates of NTM colonization in potable water systems are found in hospitals and hemodialysis and dental offices, with rates ranging from 60% to 100% [9, 10, 15–19]. Mycobacterium avium colonization is more likely in the recirculating water systems commonly used in hospitals than in nonrecirculating systems [20]. Hot water has a higher colonization rate than does cold water, and shower heads have been shown to yield the highest number of organisms [21]. In the late 1970s, an increase in the isolation of M. avium complex was noted in the northeastern United States [12, 22]. During this time, concerns regarding possible scald injuries and increasing energy costs prompted institutions to reduce hot water temperatures from 70°C to 55°C or lower [23–25]. Higher colonization rates of water systems by M. avium that resulted from reduced hot water temperatures have been suggested as a contributing factor to this increase [23]. M. avium andM. xenopi are isolated more frequently from hospital hot water sources, which reflects the optimal growth temperature of these organisms [15, 21, 26, 27], whereas M. kansasii is isolated more frequently from cold water sources [27, 28]. The microbiology of clinically significant NTM is summarized in table 1.
The Runyon classification and characteristics of clinically significant nontuberculous mycobacteria (NTM).
Molecular Methods Used in Epidemiologic Investigations
The investigation of nosocomial outbreaks of NTM infection has been greatly aided by the development of DNA typing methods. In early investigations, mycobacteria were assumed to represent a common epidemic strain if they shared the same species designation, colony morphology, and antimicrobial sensitivity profile. Phenotypic typing alone has limited discriminatory power, however, and newer methods that involve genotypic typing allow greater accuracy in epidemiological investigations [29–31]. Pulsed-field gel electrophoresis (PFGE) is a highly reproducible and powerful epidemiological tool for typing NTM; it has become a standard method for investigating nosocomial outbreaks. Guidelines have been proposed to assist the investigator in interpreting PFGE patterns and differentiating epidemic and nonepidemic strains [32].
Definitions of Nosocomial Infection and Colonization
Isolation of NTM from a clinical specimen may represent infection, colonization, or pseudoinfection. The infection may be detected during the hospital stay, but more-indolent organisms, such as NTM, can manifest after discharge [33]. The clinical presentation depends on NTM species, route of infection, and presence of underlying medical conditions [34] (table 2). “Colonization” is defined as the establishment of NTM within the patient’s microflora without evidence of disease or tissue invasion. A “pseudoinfection” is defined as a positive culture result from a patient without evidence of true infection or colonization, which is typically caused by contamination during specimen handling [35].
Association of species of nontuberculous mycobacteria (NTM) with clinical categories of nosocomial infection.
Respiratory Tract Colonization And Infection
Several outbreaks have been described in which NTM in hospital water supplies have resulted in respiratory tract colonization. Patients with preexisting chronic lung disease are at greater risk of colonization. PFGE was used to confirm that theMycobacterium fortuitum cultured repeatedly from the sputum samples of 16 asymptomatic patients in 1 outbreak was identical to the strain isolated from the hospital potable water supply. A common shower was the implicated source [36]. In a separate hospital outbreak, M. fortuitum that had colonized an ice machine was isolated from multiple serial sputum samples obtained from 30 inpatients during a 5-month period [37]. In a third outbreak, Mycobacterium chelonae that had colonized a hospital hydrotherapy pool were cultured from the sputum samples of children with cystic fibrosis who were regular users of the pool [38].
Although respiratory tract colonization by rapidly growing NTM typically does not lead to infection, more pathogenic species can cause disease in this setting. M. xenopi was isolated from numerous sputum samples and from the hospital water supply during a 7-year period in one hospital, and 19 patients developed M. xenopi pulmonary infection. Patients with chronic lung disease were at much higher risk of infection. [39]. M. xenopi caused a second hospital outbreak, in which 72 patients were colonized, which resulted in 1 case of fatal pulmonary disease. The outbreak occurred in a recently constructed hospital in which M. xenopi was isolated from 65% of potable water sources [13]. In addition to potable water, medical devices may also transmit NTM and result in infection. M. chelonae pulmonary infection occurred in 2 patients after the use of a colonized bronchoscope [40].
Infection in Patients Undergoing Dialysis
Numerous species of NTM, especially rapid growers, have caused infection in patients undergoing dialysis. These infections are typically caused by contaminated aqueous solutions or the use of colonized potable water in the processing of reusable hemodialysis filters. Nosocomial M. fortuitum infection has resulted in clusters of peritonitis cases in patients on continuous ambulatory peritoneal dialysis [41, 42]. Automated peritoneal dialysis machines colonized with an M. chelonae-like organism were responsible for 2 outbreaks of peritonitis that involved 10 patients [43]. Mycobacterium abscessus-contaminated potable water used in the cleaning of reusable hemodialysis filters resulted in disseminated disease or graft infection in 5 patients at 1 center [44]. M. chelonae was the cause of a similar outbreak among 27 hemodialysis patients when filters were contaminated during cleaning [45].
Disseminated Infection
Newer molecular techniques have been applied to the investigation of generalized NTM infections in immunocompromised hosts and have documented acquisition of the organism from the hospital water supply. One study of 36 patients with AIDS who had disseminated M. avium infection used PFGE to characterize the infecting strain. Seven patients (19%) had clustered strains of M. avium. In 5 of the 7 patients, the PFGE pattern of the clinical isolate matched that of the M. aviumstrain recovered from the hot water supply of the hospital that provided health care to these patients, indicating nosocomial acquisition of the organism [23]. A second study used PFGE to compare M. avium strains isolated from patients with and without HIV infection with those cultured from area hospital potable water systems. Several of the environmental and clinical strains were closely related, which suggests possible nosocomial infection [17]. Colonization of a hospital water system by Mycobacterium genavense has also led to nosocomial infection and disseminated disease in patients with AIDS [83].
Other types of immunosuppressed patients are also at risk. A patient with leukemia and neutropenia developed disseminated M. fortuitum infection during a prolonged hospitalization; arbitrarily primed PCR analysis of the patient and environmental isolates revealed that the hospital shower was the probable source [84]. Nosocomial acquisition of M. abscessus was likely in 2 renal transplant patients who developed disseminated infection at one hospital during a 3-month period [85].
Infections in Surgical Patients
NTM infections in surgical patients have been reported in a wide variety of settings. The use of colonized aqueous solutions and inadequate sterilization or disinfection of surgical equipment are often factors in these infections.
M. fortuitum and M. chelonae have caused multiple outbreaks of sternal wound infection and endocarditis after cardiac surgery. One outbreak was traced to contaminated tap water that had been used to make cardioplegia solution. The highest mortality rates have been reported in patients with prosthetic valve replacements [86–89]. The implantation of porcine heart valves colonized with M. chelonae during manufacture resulted in pericarditis and endocarditis [90–92].Mycobacterium smegmatis, M. chelonae, Mycobacterium gordonae, and M. fortuitum have caused isolated cases of prosthetic valve endocarditis in the immediate postoperative period [65, 93–95]. M. chelonae, M. fortuitum, and M. smegmatis have also caused isolated cases of postoperative sternal wound infections and mediastinitis [65, 96, 97]. M. chelonae has caused vein graft harvest site infections after cardiac bypass surgery [98].
M. abscessus that had colonized aqueous solutions used to mark the incision site prior to surgery has led to 2 outbreaks of wound infections [99, 100]. M. chelonaecolonization of a hospital water system resulted in 22 cases of post-rhinoplasty cellulitis due to inadequate sterilization of surgical equipment [101]. M. abscessuscolonizing potable water sources in an operating room resulted in 45 cases of postsurgical wound infections during a 5-month period. Improved sterilization techniques aborted the outbreak [102]. Rapidly growing mycobacteria have caused cellulitis and postsurgical wound infections after liposuction. The affected hospitals had used a quaternary ammonium disinfectant solution, which is an inadequate mycobactericidal agent, and tap water in the processing of the surgical equipment [103]. M. fortuitum and M. chelonae are responsible for the majority of the NTM infections seen after augmentation mammoplasty [104–107].M. fortuitum has also caused wound infections in patients after breast surgery that does not involve prosthetic implants [104]. Finally, isolated cases of M. chelonae and M. fortuitum wound infection after minor dermatologic surgery and punch biopsies have also been reported [108, 109].
M. fortuitum has caused postoperative infections after abdominal surgery. Reported cases include an abdominal wall infection after the repair of a ventral hernia that involved synthetic mesh and peritonitis in a patient who had undergone gastrectomy [110, 111].
The likely contamination by M. fortuitum of a procoagulant used in a laminectomy resulted in meningitis in one case [112]. M. xenopi spinal infections occurred in 31 patients after equipment used for arthroscopic laminectomies was liquid sterilized, then rinsed in contaminated tap water [113]. M. fortuitum has caused spinal infections after epidural injections and epidural catheter placement [114, 115]. The rare cases of ventriculo-atrial shunt infections due to M. gordonae and M. fortuitum have been attributed to wound contamination at the time of surgery [116–120].
Contaminated otolaryngostomy equipment led to M. chelonae otitis media in 17 patients [121]. M. fortuitum mastoiditis has also been reported after myringotomy and tympanostomy tube placement [122].
An unusual case of postsurgical M. avium osteomyelitis not associated with implanted prosthetic material has been reported in an immunocompetent youth. Routine mycobacterial cultures in this case were negative, and the diagnosis was made by PCR amplification of M. avium DNA sequences from clinical specimens [123].
Rapidly growing mycobacteria have caused ocular infection after corneal transplant, posterior capsulotomy, blepharoplasty, and other ophthalmologic surgery [124–128]. M. abscessus orbital infection that resulted from lacrimal duct probing has been reported [129].
Pseudoinfection
A pseudo-outbreak is a cluster of pseudoinfections and may present in a variety of ways, depending on the cause (table 3). NTM colonization of potable water systems has resulted in hospital-wide pseudo-outbreaks. Typically, an increase in frequency of NTM isolates in patients without signs of infection raises the suspicion of a pseudo-outbreak. The offending species is then isolated from multiple water sources. After implementation of control efforts, including transient hyperchlorination of the water system, decontamination of equipment, or the installation of water filters, the frequency of isolation of NTM then returns to preoutbreak levels [130–133].
Association of species of nontuberculous mycobacteria (NTM) with types of nosocomial pseudoinfection.
NTM pseudoinfection has frequently involved respiratory specimens. M. gordonaecontamination of both water used to gargle prior to sample collection and dye used during bronchoscopy resulted in 2 pseudo-outbreaks [134, 135]. M. chelonae contamination of a multiuse topical anesthetic spray led to a similar pseudo-outbreak, solved by switching to a single-use device [136].
Endoscopes, especially bronchoscopes, are frequently the cause of pseudo-outbreaks of NTM. Bronchoscope suction valves and channels are difficult to clean and disinfect, and they become colonized with mycobacteria, which may lead to the transmission of disease to previously uninfected patients [40, 137–139]. Damage to the suction channel is hard to detect, and it predisposes to NTM colonization and subsequent contamination of specimens [40]. Automated endoscope washers also result in contaminated sputum specimens, typically by NTM colonization of water-holding tanks or water inlet hoses [138, 140–146]. Manual washing can decrease the risk of bronchoscope colonization, yet a pseudo-outbreak still occurred after M. xenopi-contaminated water was used in the manual disinfection of bronchoscopes in one hospital [147]. The avoidance of tap water during endoscope cleaning helped to resolve a bronchoscope-associatedM. chelonae pseudo-outbreak [148]. Contamination of clinical specimens by NTM has resulted in unnecessary therapy directed against M. tuberculosis and delayed diagnosis of malignancy [138, 139, 145, 149].
Contamination in microbiology and pathology laboratories has also resulted in pseudo-outbreaks. An increase in positive acid-fast bacilli smears or cultures obtained from patients without a compatible clinical syndrome should prompt evaluation for a pseudo-outbreak; the pattern of isolation may suggest the source and help direct the epidemiologic investigation. NTM contamination of aqueous solutions used in mycobacterial culture has resulted in multiple pseudo-outbreaks [150–155]. Mycobacterial contamination of materials used in the transport and preparation of pathological specimens has led to false positive results for AFB [149, 156]. M. avium, M. gordonae, and M. chelonae pseudo-outbreaks have occurred after NTM colonization of the sampling needle of automated radiometric mycobacterial culture systems. Increasing the temperature and duration of the decontamination cycle of the sampling needle between specimens stopped these pseudo-outbreaks [157, 158].
Newer molecular techniques have facilitated the investigation of pseudo-outbreaks that occur during a prolonged period or involve unusual situations. Arbitrarily primed PCR has been used to investigate long-term pseudo-outbreaks caused by rapidly growing mycobacteria [159]. This technique demonstrated that a 6-year-long outbreak in a laboratory was caused by the strain of M. abscessus recovered from a distilled water source [160]. PFGE has also been used to solve pseudo-outbreaks; in one case, this technique revealed that M. avium isolated from cardiac tissue resulted after laboratory cross-contamination from a sputum sample [161].
Prevention and Control
The prevention of nosocomial infections and pseudoinfections due to NTM is challenging. Their high lipid content and triple-layered cell wall render mycobacteria more resistant to killing by disinfection, elevated temperature, and ultraviolet light, compared with other pathogenic bacteria that may colonize potable water [162–165]. In addition, NTM frequently exist within biofilms that coat internal surface pipes and fixtures of water distribution systems and storage tanks [14, 19, 166]. Biofilms appear to support mycobacterial growth and protect the organism, which makes complete disinfection of colonized water systems often difficult to achieve [141, 166–168].
Disinfectants. The effectiveness of disinfectants against mycobacteria depends on the composition and concentration of the active agent, presence of organic material, disinfectant contact time, and duration of potency. NTM, especially M. avium, are more resistant to chemical disinfection than are M. tuberculosis andMycobacterium bovis [169–172]. The ability to compare the activity of mycobactericidal disinfectants is hindered by the lack of a standardized test that is accurate, reproducible, and applicable to clinical practice [171, 173, 174]. Despite this limitation, several chemical compounds are known for their superior disinfectant activity against mycobacteria (table 4).
Mycobactericidal disinfectants.
Instruments that contact mucous membranes require high-level disinfection. Several options are effective. Glutaraldehyde is a proven and effective mycobactericidal agent in both laboratory and clinical testing and is used widely, especially in the disinfection of semicritical medical devices [174–177]. Endoscopes are reliably disinfected with 2% glutaraldehyde, if a contact time of 20 minutes at 20°C is used with an adequate precleaning and terminal rinse [176–179]. In practice, glutaraldehyde has been reused for up to 14 days in manual endoscope disinfection baths and automated washers. During this time period, glutaraldehyde concentration can decrease by up to 50%, with a subsequent weakening of mycobactericidal activity [169, 180]. One study revealed the failure of 7-day-old glutaraldehyde in a manual endoscope disinfection bath when the concentration had decreased to 1.5%. Decreases in glutaraldehyde concentration are rapid when manual disinfection baths are used more frequently and with automated washers [169]. Strains of M. chelonae have been recovered from endoscope washers that exhibit increased resistance to glutaraldehyde when compared with stock strains [181].
Agents that have also been reported to be effective endoscope disinfectants include 0.2%–0.35% peracetic acid, a 0.5% glutaraldehyde-0.03% phenolic compound mixture, and an alkyl polyguanide-alkyl quaternary mixture. Adequate contact time for these agents may be significantly less than 20 minutes [170, 172,176]. Iodophors may be used to disinfect endoscopes, but longer contact times may be required [174, 176, 177]. The potential benefits and limitations of these agents are outlined in table 4.
Ethyl and isopropyl alcohol are excellent mycobactericidal agents, but rapid evaporation limits contact time. Alcohol is very effective when it is used as the terminal rinse in endoscope disinfection [139, 142, 179, 182]. Formaldehyde was used extensively in the past but has been generally replaced by newer agents with lower toxicity [182, 183]. Iodophors are effective as a topical mycobacteriocidal disinfectant [171] but may stain certain materials [177, 182]. Hydrogen peroxide is active against mycobacteria [174], but the high concentrations needed may be potentially damaging to metals and plastics [183].
Endoscopes. Careful adherence to established decontamination guidelines are critical with endoscopy equipment. These devices are often heavily used, and their intricate design makes cleaning and disinfection difficult. Manual precleaning of endoscopes with a neutral detergent prior to disinfection is vital and results in up to a 4-log reduction in number of organisms [179]. Fundamental lapses in endoscope disinfection have been noted in observational studies, including inadequate precleaning, failure to monitor disinfectant concentration, use of outdated disinfectants, inadequate contact time, the use of tap water as a final rinse, and failure to sterilize biopsy forceps [180, 184].
Hospital water systems. Efforts to limit nosocomial infections due to NTM have also focused on controlling NTM in potable water systems. Chlorine at high concentrations (1 mg/L) is mycobactericidal, but concentrations of <0.15 mg/L are ineffective [165, 185, 186]. Although adequate levels may be present at the originating water treatment facility, chlorine levels at more distant points in the distribution system may be inadequate to inhibit NTM proliferation [186]. Transient increase in water chlorine concentration to 2–3 times baseline levels was used to successfully control colonization with Mycobacterium terrae in one hospital [131].
Many NTM may be killed or inhibited at sufficiently high hot water temperatures.M. avium and M. xenopi are thermophilic, however, and require ⩾70°C for inhibition [165, 187, 188]. This high water temperature has a significant potential for scald injury, unless safeguards are undertaken [24]. Many state regulations now limit the maximum hospital hot water temperature to 42°C–50°C [25]. Temporarily increasing water temperature to >70°C combined with flushing all faucets and showers was used to control M. xenopi colonization in one hospital [133]. Other less successful efforts to decontaminate hospital water have included the installation of filters or periodic flushing of water systems [44, 132, 141, 148,153, 168, 188].
Control of outbreaks and pseudo-outbreaks. Control of nosocomial outbreaks of NTM requires adequate surveillance by the infection control team, application of molecular typing techniques, prompt identification of the source, and institution of effective control measures. Decontamination of equipment or plumbing fixtures can be achieved by autoclaving when possible, flushing nonremovable parts with mycobactericidal agents, and adhering strictly to scheduled equipment decontamination protocols [36, 44, 45, 59, 130, 138, 141]. Modernization and preventive maintenance of hospital water systems have resulted in decreased NTM contamination of clinical specimens [189]. Sterile water should be used in the collection and processing of clinical specimens whenever possible [87, 134, 147]. To minimize the risk of cross-contamination, single-use devices and medication vials are preferable when possible [46, 47,136]. For reusable medical devices, tap water should be avoided in cleaning, and a terminal alcohol rinse is beneficial [139, 142, 148, 178, 179]. Cross-contamination in automated mycobacterial culture systems that use a specimen probe may be limited by increasing probe sterilization temperatures between specimens [157, 158]. Finally, adequate surveillance for mycobacterial contamination of medications and devices during the manufacturing process is essential [92, 93, 154].
Conclusion
NTM are ubiquitous in the environment and have the potential to cause colonization, serious infection, or disruptive pseudo-outbreaks with a wide variety of presentations. Data suggest that problems with NTM in health care facilities may be increasing, and reduced hot water temperatures may be partly responsible for this phenomenon. Attention to adequate high-level disinfection of medical devices and the use of sterile reagents and biologicals will prevent most outbreaks. Because NTM cannot be eliminated from the hospital environment, and because they present an ongoing potential for infection, NTM should be considered in all cases of nosocomial infection, and careful surveillance must be used to identify potential outbreaks. Analysis of the species of NTM and the specimen source may assist in the determination of the significance of a cluster of isolates. Once an outbreak or pseudo-outbreak is suspected, molecular techniques should be applied promptly to determine the source. Control measures should be based the NTM species, type of outbreak, and potential patient groups at increased risk of infection.
- Received July 6, 2000.
- Revision received May 14, 2001.
- © 2001 by the Infectious Diseases Society of America