© 2003 American Thoracic Society
Infection Control in the Bronchoscopy SuiteA Review of Outbreaks and Guidelines for PreventionSection of Bronchology, Departments of Pulmonary and Critical Care Medicine and Infectious Disease, Cleveland Clinic Foundation, Cleveland, Ohio Correspondence and requests for reprints should be addressed to Atul C. Mehta, M.D., Department of Pulmonary and Critical Care Medicine, Cleveland Clinic Foundation, Desk A90, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail: mehtaa1{at}ccf.org The 2000 Institute of Medicine report, To Err Is Human, suggests that medical mishaps occur in 2.9 to 3.7% of hospitalizations, resulting in 44,000 to 89,000 deaths yearly in the U.S. alone (1). In this context, March 2002 reports of bacteria transmission via bronchoscopes fostered a groundswell of concern among medical personnel and the public alike (2, 3). With nearly 500,000 bronchoscopies performed each year in the U.S. alone (4), there are fears of significant patient harm. Most bronchoscopists are familiar with a 1 to 3% rate of immediate, procedure-related complications due to bronchoscopy (5). In contrast, reports of bronchoscopic pathogen transmission have been scattered and largely anecdotal. Under-recognition and under-reporting of such episodes have contributed to a sense of complacency regarding infection control in the bronchoscopy suite. With recent events, renewed attention may be focused on infection control. However, the burgeoning sensation of time-pressure and administrative demands for economic thrift will likely intensify the temptation to "cut corners." Bronchoscopic infectious complications include distal spread of organisms within a patient during bronchoscopy, transmission of organisms to subsequent patients via contaminated instruments, accessories or solutions, and transmission of infectious agents to medical personnel or nearby patients. Distal spread of infection is rarely clinically significant, although the exact incidence is unknown (6); it may include contamination of the lower respiratory tract with organisms from the upper respiratory tract, extension of infection within the lung, and hematogenous dissemination to distant organs (6). Distal pathogen spread is a complication of the procedure itself and will not be discussed further here. This article reviews the available evidence regarding transmitted infectious hazards of bronchoscopy. We will also provide guidelines for the prevention of infections associated with bronchoscopes. PATHOGEN TRANSMISSION BY THE BRONCHOSCOPE Microbes may be spread to subsequent patients if bronchoscopes or accessories are inadequately disinfected or if contamination occurs after disinfection. Multiple avenues for contamination have been described (Table 1) . Despite this, reports of pathogen transmission via contaminated bronchoscopes have been relatively uncommon considering the large number of procedures performed. To date, the English-language literature includes at least 59 reports, totaling 953+ patients (see Tables E2 and E3 in the online supplement). Most reports have described pseudoinfections. These occur when, despite disinfection of the bronchoscope, a contaminating organism is isolated from bronchoscopic specimens in a patient with no clinical evidence of disease attributable to that organism. A pseudoepidemic or pseudo-outbreak occurs when the organism is isolated in bronchoscopic specimens from multiple patients. True infection implies development of actual clinical disease in a patient after undergoing bronchoscopy with the contaminated instrument.
Review of the available cases reveals that the data are limited in several aspects. Most of the reports are descriptive series, with few case–control investigations. In many of the episodes, the causality between the clinical scenario, putative offending organism, and ostensible mechanism of contamination is tenuous. Most importantly, because there are no prospective studies of pathogen transmission, the actual incidence is unknown. In light of the unclear significance of the problem, the urgency to intensify infection control recommendations is unclear. The most recent reports (2, 3) are the first to describe pathogen transmission despite adherence to all current (2003) reprocessing standards. Epidemiologic investigations in both instances implicated loose fittings over the valve stem for the working channel of certain bronchoscope models. In this scenario, effective mechanical cleaning and disinfection would be impossible. In total, there were multiple pseudoinfections and 20 to 43 possible true infections, including up to three deaths, due mainly to Pseudomonas aeruginosa. As both series were retrospective, it is difficult to discern the exact number of cases actually attributable to pathogen transmission; for example, DNA fingerprints from many of the clinical isolates described by Srinivasan and coworkers (2) did not match those from the bronchoscopes, and the causality for most of the true infections is unclear. However, even if a minority of the cases were actually due to contaminated bronchoscopes, the notion that they occurred in a setting where infection control guidelines were ostensibly adhered to rigorously is concerning. Presumably, there could have been many other institutions where similar pathogen transmission occurred, unnoticed or unreported. As demonstrated in Figure 1 , reports of contamination (and true infection) have become more frequent over time. Also, in the 1970s and 1980s, diverse bacteria species, such as Klebsiella and Serratia spp, were frequently implicated; since the mid-1980s, P. aeruginosa, fungi and mycobacteria have predominated. These apparent trends may reflect increased awareness, more bronchoscopies, and reporting bias. However, it is also likely that the spectrum of observed organisms has actually changed. This shift implies both that disinfection techniques have improved over time and that there are a number of organisms resistant to all but assiduous cleaning and application of powerful disinfectants. The most common mechanisms for contamination have also apparently shifted. Earlier reports frequently described use of ineffective disinfectants or lack of adequate cleaning regimens. More recently, contaminated automated endoscope reprocessors and ostensible breaches in pre-established infection control protocols based on guidelines have been identified.
Environmental or commensal organisms, especially P. aeruginosa, Serratia marcescens, nontuberculous mycobacteria, and environmental fungi are the most frequently described pathogens. A complete list of all English-language reports of pseudoinfections may be accessed in the online supplement (Table E2). True infections caused by bronchoscopy have been rare; there are 13 well-documented reports involving 21 patients in the English-language literature (Table E3 in the online supplement). Six reports also describe putative bacterial infections due to bronchoscopy, but a paucity of clinical details makes them difficult to interpret. Twenty-eight percent of all discrete reports describe at least one patient with possible true infection, but the proportion of true infections among all affected patients is only 3 to 4% (see Tables E2 and E3 in the online supplement). In general, these episodes have been detected when astute clinicians or microbiology laboratories observe dramatic or unexpected changes in isolate patterns. Attributable disease due to the contaminating organism can sometimes be traced retrospectively back to an index case. Development of molecular techniques such as DNA fingerprinting has enhanced epidemiologic investigations in recent reports (2, 3, 7–9) and should be used routinely in epidemiologic investigations of suspected outbreaks. Review of the larger outbreaks in the past suggests that effective surveillance of microbial isolate patterns is a major determinant of outbreak size; in some instances, epidemic size might have been limited by earlier discovery. Pseudoinfections may lead to unnecessary worry, unwarranted treatment of isolates, delay in diagnosing real disease, and need for further investigations to exclude infection (6). Although reporting of suspected pseudoinfections to state health departments and the Food and Drug Administration is highly encouraged, they are likely under-reported (10). Another potential type of pseudoinfection that may be increasingly seen in the future is the false-positive polymerase chain reaction assay because improperly cleaned bronchoscopes may harbor amplifiable nucleic acid residues (11). BRONCHOSCOPE REPROCESSING Table 2 outlines terms frequently used in disinfection and sterilization practices, as devised by Spaulding (12). There are problems specifically associated with bronchoscopes using this scheme to classify disinfection practices. Is the bronchoscope still a semicritical item when it is used with sterile biopsy forceps or used in a patient with pulmonary hemorrhage? Although sterilization would be most effective in preventing pathogen transmission, flexible bronchoscopes are not easily sterilized. Ethylene oxide gas is effective but requires 24 hours, and luminal penetration may be incomplete if cleaning is inadequate (13). Steam sterilization is quick but damages the fiber-optic parts. From a practical standpoint, sterilization is unnecessary: it is costly, time-consuming and unlikely to confer significant additional protection (14). A summary of recommendations for reprocessing is provided in Table 3 .
Factors impeding disinfection are inherent in bronchoscope design: long, narrow working channels (2.6 mm diameter), multiple internal angulations, mated surfaces, springs, and valves. Recent episodes highlight the fact that even careful instrument design may not completely obviate these characteristic risks (2). Effective disinfection necessitates mechanical removal of any biofilms and of encrusted organic material (15). Organic debris may preclude contact of the disinfectant with micro-organisms or even inactivate some agents (15). Even gas sterilization may not be effective in these circumstances (13). This is not a trivial problem because immediately after routine bronchoscopy, instruments are contaminated with an average of 6.4 x 104 cfu/ml of bacteria (16).
Cleaning Technique
Disinfection All currently approved agents are effective high-level disinfectants in experimental conditions, by definition achieving a greater than 4 log reduction in microbial burden (14). Parenthetically, meticulous cleaning alone achieves a 3.5 to 4 log reduction in organism load (23). The choice of specific disinfectant will likely vary by institution, depending on cost, volume of procedures, use of automated endoscope reprocessors, number of bronchoscopes, and cleaning facilities available. For the practical purpose of achieving high-level disinfection, the choice of agent is probably less important than careful cleaning and assiduous adherence to an appropriate reprocessing protocol (14). A complete listing of Food and Drug Administration–approved disinfectant formulations, conditions, and maximum reuse time can be accessed online (24). Dilution of solutions to ineffective levels may occur more quickly than the maximum reuse period, however, necessitating periodic testing of solution concentration and pH with commercially available test kits (25). Glutaraldehyde solutions should not be used at concentrations less than 2%. Even with adequate concentration, disinfectants should not be used beyond the listed time frame because spontaneous polymerization of aldehyde groups over time abrogates their microbiocidal activity (21).
Special considerations with automated bronchoscope reprocessing. Automated endoscope reprocessors help maintain standards of disinfection, ensure consistency between operators, and eliminate human errors. User manuals should be easily accessible and should provide information on which specific bronchoscope models have been tested for compatibility (6). For example, failure to provide adequate internal channel penetration of disinfectant stemming from use of wrong connectors has caused outbreaks (9). Biologic and chemical markers are available to assess disinfectant strength, pH, and efficacy of decontamination. Provided adequate cleaning and high-level disinfection are achieved for all devices, bronchoscopists can use the same automated reprocessors that are used for gastroenterologic endoscopes.
Disinfection of hepatitis B, hepatitis C, human immunodeficiency virus, or mycobacteria. The main theoretical risk of virus transmission resides in potential failure to adequately remove biologic debris via mechanical cleaning, thus allowing viruses to escape contact with disinfectants. Failure to clean endoscopes adequately has been shown to preclude effective disinfection of hepatitis B and C viruses (28). Although some authors have suggested intensified disinfection regimens (or sterilization) after bronchoscopy of patients with known viral infections, there is no evidence to support this recommendation. Because serologic status is usually unknown, it is more prudent to simply use adequate precautions for all patients. As described earlier, mycobacteria have been responsible for a large proportion of reported contamination episodes. All historical tuberculosis (TB) cases have been associated with at least one lapse in standard infection control procedures; thus, it is reasonable to infer that they may have been preventable. Given the absence of supportive case–control studies, however, the level of evidence for this conclusion is low. Disinfection is difficult, especially for nontuberculous mycobacteria because they are both environmentally ubiquitous and the most difficult micro-organisms to eradicate, excepting bacterial endospores (32). Some species (Mycobacterium xenopi, Mycobacterium avium) tolerate hot water temperatures and are capable of slow growth even in distilled or chlorinated water (33). Although some authors have advocated longer disinfection times following patients with mycobacterial disease or preceding immunocompromised patients, this strategy is unnecessary if all current infection control guidelines are followed (14). Numerous studies have demonstrated that 20 minutes in 2% alkaline glutaraldehyde at 20°C or periacetic acid provides adequate disinfection if cleaning with detergent precedes disinfection (14, 15, 23, 34). Experimental models suggest that even high-level contamination (108 cfu/ml of Mycobacterium gordonae) is completely eliminated by this regimen (34).
Uncommon infectious agents. To date, a total of 23 cases of anthrax have been reported to the CDC following the intentional distribution of Bacillus anthracis spores through the U.S. postal system. Eleven cases were inhalational anthrax with five deaths, of which at least one patient underwent a bronchoscopic procedure (35), raising the issue of disinfection. B. anthracis is a large, spore-forming encapsulated aerobic, nonmotile toxin producing gram-positive rod. Although anthrax spores are resistant to high-level disinfectants, they are only produced in soils and in dead tissues, not in blood or living tissues. High-level disinfection of bronchoscopes is therefore adequate for patients with known or suspected inhalational anthrax.
Postdisinfection Handling Instruments should be stored in an upright (hanging) position to prevent accumulation of moisture, and valves should not be reassembled until the next procedure (36). Outbreaks have occurred by both of these mechanisms (37). Other accessories should be adequately disinfected or sterilized. For example, up to 75% of atomizer lumens and 42% of their fluid reservoirs may be contaminated after a single use (38). Active TB has reportedly developed by this mechanism (8). Reusable instruments, such as biopsy forceps, should be cleaned and sterilized because they penetrate intact mucosa during normal use. Cleaning may be difficult due to multiple crevices and tightly wound coils. Ultrasonication is most efficacious for cleaning these devices (16). PREVENTION OF OUTBREAKS Bronchoscopists or associated microbiology laboratories should routinely review pathogen isolates to identify unexpected clusters or trends. Settings with fewer procedures, multiple bronchoscopists sharing the same equipment, no staff or physical area dedicated solely to bronchoscopy, and absence of routine training in disinfection techniques are at higher risk of missing such trends and therefore require especially well-formulated procedures for carrying out isolate surveillance. If investigation of suspected episodes is needed, cultures should include samples from all internal lumens of the bronchoscope, internal surfaces of automated endoscope reprocessors, and tap water if used for rinsing. Brush cultures of the bronchoscope lumens are more sensitive than saline flushes (14). The issue of periodic surveillance cultures of the bronchoscope, accessories, automated endoscope reprocessors, or tap water supplies remains controversial. Further study is needed in this area. Specific issues which are unclear at present include how often to culture, what items specifically to culture, the fiscal implications of such a strategy, what type of organisms are considered significant and in what quantities. Until evidence is available, we do not recommend routine surveillance cultures. To our knowledge, until 2003 (2, 3), nosocomial infections had not been reported in any case where all current guidelines were followed carefully. However, adherence to published preventive guidelines is poor (15, 39, 40). For example, an observational study of 26 U.S. facilities revealed that the vast majority of endoscopes and bronchoscopes were improperly disinfected (39). Procedural breaches included omission of any disinfectant, failure to routinely test disinfectant concentration, not cleaning or flushing all ports, failure to time manual disinfection periods, and not fully immersing the endoscope in disinfectant solution. In 78% of facilities, biopsy forceps were not sterilized after each use. As a result, washings from 17 of 71 (23.9%) of gastrointestinal endoscopes were culture-positive for more than 105 bacterial cfu/ml. A more recent audit of practice in the UK was equally disturbing—the vast majority of centers did not follow national guidelines (40). In this survey, for example, 43% of departments did not rinse bronchoscopes with sterile or filtered water after chemical disinfection, despite numerous literature reports of microbial tap water contamination. Hospitals with dedicated endoscopy units and staff training sessions complied more closely with standards. TRANSMISSION TO PERSONNEL AND BYSTANDERS
Risks of Diagnostic Bronchoscopy Infection precautions should be in place for every procedure, including full barrier clothing (gowns, gloves, masks, and eye shields), needlestick precautions, and adequate ventilation. The bronchoscopy area should have engineering controls that will allow for negative air pressure, at least 14 air changes per hour, and either discharge of air directly to the outside or monitored high-efficiency particulate air filtration before recirculation. A 1997 survey revealed that 93% of bronchoscopists did not routinely wear protective clothing (40). This finding is alarming because there are well-documented instances of human immunodeficiency virus transmission through mucocutaneous contact with blood and body fluids (45). For patients with suspected or confirmed TB infection, the need for bronchoscopy should be carefully weighed against the risks to staff and bystanders. Where available, we recommend use of a power air-purifying respirator hood, which provides superior protection (6, 46). The N95 particulate respirator is a minimally acceptable alternative (46). Needles should not be used to remove biopsy specimens from forceps: hepatitis B has been transmitted by an accidental stick during this maneuver (29).
Risks of Therapeutic Bronchoscopy
Conclusion Specific approaches to improve infection control practices include formalized institutional monitoring of isolate patterns to detect trends early, predefined protocols for epidemiologic characterization of suspected outbreaks, and institution of quality control monitoring for all steps in instrument reprocessing. Molecular biology techniques, such as DNA fingerprinting should be used where available when outbreaks are suspected. In contrast, there is no proven role for routine surveillance cultures at present. Quality control for reprocessing procedures and compliance with universal precautions, including monitoring adherence to proper technique, may be especially needed. Finally, design and engineering of bronchoscopes, accessories and reprocessing devices should be attentive to infection control principles, with a focus on streamlining the process of appropriate instrument reprocessing. FOOTNOTES This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form August 2, 2002; accepted in final form February 4, 2003 REFERENCES
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
