Data Suggesting that the Practice of General Anesthesiology and Anesthesia Equipment Are not Source of Infection
Some investigators maintain that anesthesia machines, even when contaminated, do not transmit significant numbers of bacteria because (a) microorganisms are unlikely to survive in the hostile environment of the anesthesia machine due to desiccation by the flow of cold, dry anesthetic gases
(54) and (b) the rubber and metal parts of the machine and the highly alkaline condensate at the bottom of the CO
2 absorber inhibit growth of bacteria
(55,
56). The literature discussed below provides evidence in support of these hypotheses.
Colonized or Infected Patients Are not Likely to Contaminate the Anesthesia Machine Several studies suggest that patients colonized or infected with bacterial pathogens rarely contaminate the anesthesia machine
(55,
57,
58). For example, Du Moulin and Saubermann
(55) studied 15 patients anesthetized with sterile machines. Two throat and sputum cultures were obtained from each patient before general anesthesia was administered; 40% of the patients had cultures yielding more than 10 colonyforming units (CFU) of gram-negative bacteria and 60% had cultures that did not yield gram-negative bacteria. The investigators isolated 1 to 9 CFU per segment of the breathing circuit. However, all cultures from machines used on colonized patients were negative and only three cultures from machines used on patients without gram-negative colonization were positive. Similarly, Stemmermann and Stern
(57) asked 14 patients with cavitary tuberculosis to breathe into a basal metabolic rate machine (which is similar to an anesthesia breathing circuit) for 10 minutes and then cough. The investigators did not identify
Mycobacterium tuberculosis in smears or cultures of the saline used to wash the masks and tubing and, thus, concluded that the anesthesia circuit was unlikely to contribute significantly to bacterial
crosscontamination. These data suggest that colonized or infected patients rarely transmit bacterial pathogens via the breathing circuit to the anesthesia machine.
Contaminated Anesthesia Machines Are Unlikely to Transmit Bacterial Microorganisms to Patients Several investigators have used simulations to show that contaminated anesthesia machines are unlikely to transmit bacterial pathogens
(55,
59,
60,
61 and 62). For example, Adriani and Rovenstine
(59) were unable to grow microorganisms from air blown through soda lime canisters contaminated with large numbers of
E. coli or
M. tuberculosis. Ziegler and Jacoby
(60) used a contaminated machine to ventilate a sterile reservoir bag for 30 minutes; cultures of the reservoir bag remained negative. After inoculating the expiratory port of a sterile circle system with 10
8 to 10
9 CFU of either
Enterobacter cloacae or
Flavobacterium species, du Moulin and Saubermann
(55) blew 3 L of nitrous oxide and oxygen per minute through the valve for 3 hours. Every 30 minutes, they obtained samples from the valve and found progressively fewer bacteria in the cultures. They did not recover the indicator microorganisms from other parts of the machine. Ibrahim and Perceval
(62) seeded cleaned circuits with viridans streptococci or staphylococcal bacteriophages, attached these circuits to machines, and blew air through them. Air sampling cultures obtained from the distal ends of the tubes were all negative.
Anesthesia circuits are not a major source of infections for surgical patients and bacterial filters are not an important preventive measure.
Two prospective clinical trials have been cited as evidence that anesthesia circuits are not a major source of infections for surgical patients. Garibaldi et al.
(63) randomly assigned 257 patients to be anesthetized with disposable corrugated plastic circuits containing bacterial filters (0.22 µm) and 263 patients to be anesthetized with disposable corrugated plastic circuits without filters. The postoperative pneumonia rates for the two groups did not differ significantly, but the study was powered to detect a 50% difference in rates, which would be difficult to achieve with most interventions. Feeley et al.
(64) found no difference in postoperative pneumonia rates between 138 patients anesthetized with sterile disposable circuits and 155 patients anesthetized using clean reusable circuits. However, the study had a power of only 17% to detect a 50% difference in rates.
Van Hassel et al.
(65) reviewed 9 years of surveillance data and found lower respiratory tract infections in 5 of 2,300 (0.2%) patients undergoing operations under regional anesthesia and 31 of 23,500 (0.1%) patients undergoing general anesthesia with tubing that was cleaned only once a day (i.e., was shared by three to seven patients). They changed the soda lime every 3 days and they placed filters at the T pieces only when patients had suspected or overt respiratory tract infections,
M. tuberculosis, or human immunodeficiency virus (HIV) infection. They concluded that “in our setting, patient factors are most important in the development of postoperative lower respiratory infections and that the role of bacterial filters as a preventive measure is negligible”
(65).
In summary, the studies reviewed in this section suggest the following: (a) colonized or infected patients are not likely to contaminate the anesthesia machine; (b) the anesthesia machine does not serve as a significant reservoir for bacterial microorganisms; (c) anesthesia circuits are not a major source of infections for surgical patients and bacterial filters are not an important preventive measure.
Data Suggesting that the Practice of General Anesthesia, Anesthesia Equipment, the Operating Room Environment, Anesthetic Medications, and Anesthesia Personnel Are Associated with HAIs
Endotracheal Tube Studies in experimental animals indicate that endotracheal tubes disrupt the ciliated tracheal epithelium, cause an inflammatory response, and impair mucociliary clearance, increasing the risk of subsequent infections
(66,
67,
68,
69,
70 and 71). In addition, the endotracheal tube can serve as a route by which bacterial pathogens from the patient’s oropharynx, healthcare providers’ hands, and the surrounding environment can be transmitted into a patient’s trachea
(72,
73). Indeed, several studies have shown that gram-negative bacteria or other potential bacterial pathogens that were not identified by preoperative cultures of the nasopharynx, pharynx, or larynx subsequently contaminated endotracheal tubes in the postoperative period
(72,
73 and 74). In addition, nasal-tracheal intubation has been shown to cause transient bacteremia
(75,
76). Thus, placement of endotracheal tubes can increase the risk of transmitting bacterial pathogens to patients and can contaminate a patient’s lungs with his or her own flora, thereby increasing the risk of HAIs.
Anesthesia machine and circuit (Ambu bag, breathing circuit tubing, Y connector, inspiratory and expiratory valves, CO2 absorber): Numerous authors cite an outbreak of follicular tonsillitis
(77) and two outbreaks caused by
Pseudomonas aeruginosa (78,
79) as evidence that contaminated anesthesia machines transmit microorganisms. These reports provide some evidence that anesthesia equipment (anesthesia machine, Ambu bag, and circuit) could be reservoirs for bacterial pathogens and may play a role in bacterial transmission. However, the results of these investigations should be interpreted with caution, because the authors did not conclusively identify the source of the infecting microorganisms. Furthermore, the typing methods evaluated phenotypic, not genotypic, characteristics that do not discriminate between strains and modern molecular typing methods.
Albrecht and Dryden
(80) also evaluated whether sterilizing the breathing circuits and using a disposable absorber affected the rate of postoperative pneumonia. They retrospectively reviewed the medical records of 220 randomly selected patients who underwent major abdominal operations requiring general anesthesia and who were not infected at the time of the operation. Twenty percent (10/50) of patients who underwent operations before the anesthesia equipment was sterilized, 26% (13/50) of patients who underwent operations when breathing circuits were sterilized but the absorbers were reused, and 6% (7/120) of patients who underwent operations after sterile breathing circuits and disposable absorbers were used acquired postoperative pneumonia. The investigators concluded that contaminated anesthesia machines transmitted bacteria to patients and they should, therefore, be
sterilized between cases. However, the results of this study should be interpreted with caution, because it was an unblinded, uncontrolled retrospective study and because the investigators did not specify their definition of postoperative pneumonia.
Other investigators have evaluated the ability of microorganisms to survive on anesthesia equipment. Investigators either obtained cultures from anesthesia equipment after routine use or conducted
in vitro studies. In general, such experiments identified a wide variety of bacteria (saprophytic micrococci and bacilli,
Staphylococcus aureus, coagulase-negative staphylococci,
E. coli,
P. aeruginosa, viridans streptococci,
Bacillus species) contaminating all parts of used anesthesia machines (breathing circuits, rebreathing bags, inspiratory and expiratory valves) with the parts of the machine closest to the patient most heavily contaminated
(78,
81,
82,
83 and 84). For example, Meeks et al.
(84) obtained cultures from anesthesia equipment after use. They grew
Staphylococcus epidermidis from 73% of face masks, 12% of Y connectors, 6% of breathing circuits, and 6% of rebreathing bags;
S. aureus from 10% of face masks and 1% of rebreathing tubes; and
Pseudomonas species from 36% of face masks, 67% of Y connectors, 42% of breathing circuits, and 79% of rebreathing bags. In the study by Livingstone et al.
(85), 33% (13/39) of the rubber masks used to anesthetize patients with tuberculosis yielded
M. tuberculosis. Likewise, several investigators who obtained cultures from face masks used to administer nitrous oxide for dental procedures demonstrated that bacteria from a patient’s nose and mouth contaminated the apparatus
(86,
87 and 88).
Investigators have also investigated the potential filtering and bactericidal roles of soda lime canisters and found that they do
not filter bacterial microorganisms effectively
(83,
89,
90). For example, Murphy et al.
(89) aerosolized eight different bacterial species into a soda lime canister and found that up to 40% of the microorganisms were not retained in the canister. Investigators have also demonstrated that the soda lime in the canister is
not uniformly bactericidal. Murphy et al.
(89) found that, at room temperature, 1 gram of soda lime killed
Klebsiella pneumoniae, Candida albicans, S. aureus, P. aeruginosa, Serratia marcescens, E. coli, and
S. pneumoniae within 10 minutes, but 1% of the
Bacillus subtilis CFU survived at 30 minutes. Dryden
(83) demonstrated that 4% sodium hydroxide, Sodasorb extract, and Baralyme extract killed
P. aeruginosa and
Proteus mirabilis within 15 minutes, but
M. tuberculosis survived for at least 3 hours in each of the solutions.
Investigators have also used laboratory models to simulate the patient-anesthesia machine interaction and concluded that air could move bacterial pathogens through the breathing circuit
(83,
90,
91). For example, Nielsen et al.
(91) measured the bacterial content of anesthetic gases before and after passing them through clean and previously used breathing systems to determine whether anesthetic gases could become contaminated when blown through contaminated circuits. Gases passed through clean circuits contained 1.2 to 50.2 (median 4.2) CFU of bacteria per 100 L, compared with 3.3 to 129.8 (median 38.5) CFU of bacteria per 100 L for gas passed through used circuits (Mann-Whitney
p < .01). The authors concluded that anesthetic gases can transfer microorganisms.
Investigators have also attempted to assess the origin of breathing circuit contamination. Rathgeber et al.
(92) obtained cultures of breathing circuits used with filters and from those used without filters to assess the origin of microbiologic contamination. When a filter was used, the microorganisms isolated from the breathing circuits were different than the microorganisms detected in the patients’ tracheal aspirates. When filters were not used, the same microorganisms were isolated from the patients’ tracheal aspirates and from the tubing in 13% of the cases. Thus, the study results demonstrated that patients’ bacterial microorganisms can contaminate the breathing circuits and that filters can prevent circuit contamination. However, the investigators were unable to show that filter use changed patient outcomes, because patients were not followed prospectively to determine whether they acquired postoperative pneumonia.
Investigators from the New South Wales (NSW) Health Department studied the potential role of anesthetic equipment in intraoperative viral transmission when they evaluated a cluster of patients who acquired hepatitis C virus (HCV) infection after having operative procedures at a private hospital in Sydney
(93). After two persons who had operations on the same day presented to the hospital with acute hepatitis C, NSW health officials tested all patients who had operative procedures during the same session. Three more patients were found to be anti-HCV positive. Surgical personnel were tested and were anti-HCV negative. Patient-to-patient transmission was likely, because all five patients were infected with hepatitis C of the same genotype. The common denominator between patients seemed to be the anesthesia equipment; the same anesthesia circuit was used without a filter and without decontamination for all 11 patients who had procedures during the implicated session. On the basis of these data, the investigators concluded that the HCV was transmitted through a contaminated anesthesia circuit. They hypothesized that the index case’s respiratory secretions containing HCV were introduced into the anesthesia circuit and that the virus was transmitted in droplets through minor breaks in the oropharyngeal mucosa of subsequent patients. In response, NSW health officials recommended enforcing existing guidelines that a filter be used in the anesthesia circuit to prevent cross-transmission
(18,
19,
94).
A number of other agencies, including the AANA
(8), the Blood-borne Viruses Advisory Panel of the Association of Anaesthetists of Great Britain and Ireland
(22), the Department of Health of the Netherlands Committee on Infection Prevention
(23), and the Societé Francaise d’Anesthesie et Reanimation
(24), have recommended that an appropriate filter be placed between the patient and the breathing system and that either a new filter or a new breathing circuit should be used for each patient. At present, there is no consensus on whether hydrophobic pleated membrane filters are necessary or whether electrostatic filters are adequate. Most studies of filtration efficiency have indicated that the hydrophobic filters are more efficient
(95,
96 and 97). However, a study of patients undergoing general anesthesia found no difference between hydrophobic filters and electrostatic filters
(98). Both filter types significantly decreased the incidence of bacterial contamination in the breathing circuits compared with the level of contamination in endotracheal tubes.
Laryngoscopes Inadequate disinfection of laryngoscope blades and handles has been associated with clusters of infection
(99,
100). These clusters are discussed further in the section “Current Infection Prevention and Control Guidelines and Current Anesthesia Practice.”
Equipment in the Anesthesia Work Area Loftus et al. (
101,
102) and Koff et al. (
103) have documented that various pieces of equipment in the anesthesia work area are contaminated with bacterial pathogens. Other investigators have found extensive blood contamination in the anesthesia work area
(104). These findings are discussed further in the section “Exposure of Anesthesia Personnel to Patients’ Blood and Body Fluids.”
Air Air in operating rooms can become contaminated with bacteria. For example, Edmiston et al.
(105) obtained air samples from a single operating room during 70 different vascular procedures.
S. aureus and various coagulasenegative staphylococcal species were recovered from 64% and 86% of all samples, respectively. Gram-negative bacteria were recovered less frequently (33%). The magnitude of contamination increased with proximity to the surgical field. Some of the microorganisms were identical to those recovered from HCWs’ nares, suggesting that the surgical masks were inefficient
(105).
Medications In addition to impairing host defenses, anesthetic medications can become contaminated with viral or bacterial pathogens, which can then be injected directly into the patient’s intravascular space
(106,
107,
108,
109 and 110). This has occurred when syringes become contaminated during use, when anesthesia personnel contaminate anesthetic medications either by contaminating multidose vials or by handling medications, such as propofol, improperly
(106,
107,
108,
109,
110 and 111). The role of contaminated medications is discussed further in the sections “Infections Associated with Intravenous Anesthesia and Outbreaks Associated with Anesthesia Personnel.”
Anesthesia Providers Anesthesia providers can be colonized or infected with pathogens that can be transmitted to patients
(112,
113,
114,
115,
116,
117,
118 and 119). In addition, anesthesia providers’ hands are often contaminated with pathogenic bacteria during all phases of anesthesia: induction, maintenance, and emergence (
120,
121). The role of microorganisms colonizing or infecting anesthesia providers and microorganisms carried on anesthesia providers’ hands is discussed further in the sections “Infections Associated with Intravenous Anesthesia and Outbreaks Associated with Anesthesia Personnel.”
In summary, the studies reviewed in this section suggest that (a) Bacteria can contaminate all parts of anesthesia circuits, but the highest numbers of bacteria contaminate the parts closest to the patient; (b) Anesthetic gases may carry bacteria from the machine to the patient or vice versa; (c) The soda lime removes bacteria imperfectly and, although it kills many bacterial pathogens, M. tuberculosis and Bacillus species survive prolonged exposure; (d) Filters decrease contamination of breathing circuits; (e) Anesthesia equipment such as endotracheal tubes, the anesthesia machine, laryngoscope handles/blades, syringes, and medications can serve as reservoirs for bacterial pathogens and may facilitate transmission to patients.
Thus, there are several ways that anesthesia practice could facilitate transmission of bacteria to patients during surgical procedures
(49,
50,
51,
52 and 53). Given that surgical patients often have multiple comorbidities
(47) and that numerous host defenses are breached or impaired by the surgical incision and by general anesthesia, patients may be particularly susceptible to microorganisms transmitted in the operating room
(48). This hypothesis is supported by the results of a prospective, observational study by Hajjar and Girard
(122). They found an incidence of 3.4 HAIs per 1,000 patients during the first 72 hours after the operations, suggesting that the source of the infections may have been in the operating room
(122). However, the investigators could not directly link the practice of anesthesia or anesthesia equipment to intraoperative transmission of bacterial pathogens to patients. Consequently, many anesthesia providers do not believe that anesthetic practice or the anesthesia work area is associated with HAIs
(123). In fact, the studies reviewed thus far provide little objective evidence linking either the practice of anesthesia or anesthesia equipment with direct transmission of bacterial microorganisms to patients.
Recently, Loftus et al. (
101) developed and validated a method for assessing intraoperative bacterial transmission to the anesthesia work area and to the stopcocks on the patients’ intravenous catheters. These investigators randomly selected 61 operating rooms and decontaminated the adjustable pressure-limiting (APL) valve complex and the agent dial (AD) before the first case of the day. After the case, they cultured the APL valves, the ADs, and the stopcock sets. The number of CFU per surface area sampled (CPSS) on the APL valves and the ADs increased significantly, and 32% (95% confidence interval [95% CI] 20.6-44.9%) of the stopcock sets became contaminated. Most of the contaminating bacteria were skin microorganisms, but these microorganisms can cause bloodstream infections. In addition, methicillin-resistant
S. aureus (MRSA) was transferred to the APL valves for two patients; the stopcock set became contaminated with
E. cloacae for one of these patients. VRE was transmitted to all three sites for one patient and pulsed-field gel electrophoresis documented that all three sites were contaminated by the same strain. Moreover, the probability that the stopcock set would become contaminated increased as the CPSS increased, even after adjusting for the CPSS at baseline and for covariates (odds ratio [OR] 1.67; 95% CI 1.10-2.53;
p = .02).
Subsequently, these investigators extended their observations by assessing transmission of bacteria in 82 pairs of patients (i.e., the first and second cases done in 82 randomly selected operating rooms during the study days) (
102). The investigators also cultured the dominant hands of the anesthesia providers before they touched the patients. The investigators used biotyping to determine whether microorganisms cultured from the anesthesia work area (i.e., APL valves and ADs) and from the anesthesia providers’ hands were the same. Loftus et al. found that 11.5% of the stopcocks became contaminated, of which 47% were contaminated with isolates found on the anesthesia providers’ hands. They identified intraoperative transmission to the anesthesia work area in 89% of the cases and 12% of these work areas were contaminated with isolates from the providers’ hands. In one instance, they found
the same microorganism on the hands of the anesthesia provider before the start of the first case, on the stopcock at the end of the first case, on the anesthesia machine at the start of the second case, and on the stopcock and the machine at the end of the second case, suggesting that the anesthesia provider did not perform adequate hand hygiene and that the machine was not cleaned adequately between cases. Most transmission events involved coagulase-negative staphylococci (
n = 8), or
Micrococcus spp. (
n = 5), but a
Streptococcus spp. (
n = 1), methicillin-susceptible
S. aureus (
n = 1), MRSA (
n = 1), and
Pseudomonas spp. (
n = 2) were also transmitted. Given their methodology, the investigators felt that these percentages were minimal estimates of the actual transmission rates from the anesthesia providers’ hands. Furthermore, they found that the number of rooms that attending anesthesiologists supervised simultaneously was an independent predictor of transmission events that could not be linked to providers, suggesting that the attending physicians may have transmitted microorganisms from one patient to another. The investigators also found that patients discharged from the operating room to an intensive care unit (ICU) had a higher incidence of transmission events that could not be linked to providers, suggesting that anesthesia providers may have omitted hand hygiene, because they thought they needed to expedite care for more seriously ill patients. Again, given the limitations of their methods, the investigators felt that their results represented a minimum estimate of the transmission events.
Koff et al. (
103) extended this model further when they did an intervention to see whether increasing hand hygiene decreased environmental contamination. They significantly increased (27-fold;
p < .002) hand hygiene adherence over the baseline rate by giving anesthesia providers dispensers for alcohol-based hand rub that could be attached to their clothing. In addition, they noted that contamination of the anesthesia work area and stopcocks decreased from 32.8% to 7.5% (OR 0.17; 95% CI 0.06-0.51;
p < .01) and that the incidence of HAIs decreased from 17.2% to 3.8% (OR 0.19; 95% CI 0.00-0.81;
p = .02).
Conclusions Regarding the Role of the Practice of General Anesthesiology and Anesthesia Equipment as Potential Sources of Infection Until recently, the clinical importance of microorganisms isolated from anesthesia machines and their role in postoperative infections had not been clearly defined. In fact, Hogarth
(124) concluded following a thorough review of the available literature that there was little evidence to implicate anesthesia machines and breathing systems as either a source of pathogenic bacteria or a vector for transmitting these microorganisms to patients undergoing general anesthesia for surgical procedures. Even the outbreaks reported by Joseph
(77), Tinne et al.
(78), and Olds et al.
(79), and the report by Chant et al.
(93) provide little evidence for transmission of pathogens by anesthesia machines or equipment, because these studies did not use sensitive methods for identifying specific strains and because they did not address whether anesthesia providers complied with critical infection prevention and control practices, such as performing hand hygiene, changing gloves between procedures on the same patient, changing gloves between patients, and cleaning and disinfecting the anesthesia cart and equipment between cases
(125). Furthermore, most investigators assessing the role of the anesthetic equipment and staff in the transmission of microorganisms used simulations and did not assess real-life anesthetic procedures in operating rooms. Thus, prior studies do not allow us to determine whether the providers, the patients, or the anesthesia equipment was the source of the infecting microorganisms.
The studies by Loftus et al. (
101,
102) and Koff et al. (
103) were the first to demonstrate that anesthesia environment
and anesthesia providers
do transmit bacteria to patients and to document that increasing hand hygiene decreases transmission of bacteria to the anesthesia work area and to stopcocks on patients’ intravenous catheters. In addition, their work suggests that transmission of microorganisms in the operating room may not be benign in that the mortality rate was higher for patients whose stopcocks became contaminated (
101) and HAI rates were higher in the preintervention period when hand hygiene was poor and environmental contamination was high (
102). While their studies did not prove the direct link between poor hand hygiene in the operating room and poor patient outcomes, these studies describe a method that other investigators can use to further this work and they provide a rationale for implementing interventions. Moreover, anesthesia providers can incorporate these interventions easily into their work flow to improve hand hygiene.
Preventing Infections Associated with General Anesthesia Procedures Current recommendations for measures for preventing intraoperative transmission of pathogenic microorganisms are relatively sparse (
Table 60-1). We support changing and/or disinfecting breathing circuits and masks between operative cases
(7,
8). Because in-line circuit filters effectively prevent transfer of bacteria from the patient to the anesthesia machine and from the machine to the patient
(126,
127 and 128), we think filters should also be used routinely, particularly for patients with active pulmonary tuberculosis receiving general anesthesia
(29). Hand hygiene, cleaning, disinfection, and sterilization of equipment, and environmental cleaning are also important preventive measures. Further studies are needed to identify additional sources of pathogens in the operating room and additional risk factors for intraoperative transmission. Such studies could provide the evidence base for implementing intraoperative preventive measures.
