William R. Jarvis
Concern about the infection hazard of environmental microorganisms arises because of our close interaction with the environment and its high content of microorganisms, including important human pathogens. Microbes are remarkably efficient at becoming dispersed to virtually all unprotected sites. Where there is moisture and organic material, proliferation to large numbers occurs. Even on dry, infertile surfaces, microbes survive in various relatively inactive states. Unfortunately, although it is easy to establish the presence of microorganisms in the environment, it is difficult to assess their role in causing human disease. Evaluating the evidence in this matter is the fundamental subject of this chapter.
The scope of this chapter is determined partly by logic and partly by tradition. The title itself is a misnomer: If the environment were inanimate, it would not require discussion. The major focus is on those normally nonsterile items that may serve as fomites, or vectors of infectious agents [1]. The word fomite, although in disfavor among some researchers, remains quite useful. A fomite is an inanimate object that may be contaminated with microorganisms and serve in their transmission. The origin of the term fomites is the Latin plural of omes, the genus of fungus that was used as tinder. The dried fungus is porous and, thus, was considered “capable of absorbing and retaining contagious effluvia.” Consideration of this topic is extended to items that often are sterilized even though the need to do so is arguable (e.g., the internal surfaces of respirator and anesthesia breathing circuits, water in humidifier reservoirs [2], endoscopes that are to be passed through nonsterile cavities, or reusable pressure transducer heads) [3]. The chapter does not discuss items that clearly must be sterile. Some inclusion distinctions are quite arbitrary. Pus while in a patient's wound or on the unwashed hands of a healthcare worker (HCW) would not be considered part of the inanimate environment, but as soon as it is deposited on a surface, it would. Skin is not considered in this chapter, but airborne squamae are. Finally, elements traditionally considered environmental also are included, such as potted plants, cut flowers, insects, and problems associated with animal visitation in the hospital. This chapter also discusses the ultimate concern about the inanimate environment, ultraclean protective environments for immunosuppressed patients, disinfection and sterilization, and routine microbiologic monitoring of inanimate objects in the hospital.
The simplistic dichotomous view that we can apply to sterilization (i.e., an item is sterile or it is not) is not applicable to the factors under discussion. Instead, we must contend with the difficulty of determining the appropriate degree of contamination. In some instances, standards exist (e.g., dialysis water), but, for the most part, no standards have been set, and there is little rational basis for setting them. Setting such standards is made more difficult because microbiologic classification of environmental organisms is not as well developed as it is for organisms from clinical specimens.
There are paradoxes associated with microbiologic content of the environment. Some objects may be sterile as a by-product of manufacture. For instance, the protected outside surfaces of intravenous bottle stoppers usually are sterile, although manufacturers do not make the claim of sterility because it is burdensome to prove it to regulatory
P.276
agencies. There are items that are not marketed as sterile but that usually are (and arguably should be). When contaminated, these items have been responsible for outbreaks including karaya ostomy bags or elastoplast, which have caused Rhizopus spp. [4,5] and Clostridium perfringens [6] skin infections; contaminated blood collection tubes have caused pseudobacteremias [7,8] and true bacteremias; and contaminated hand-care products have been implicated as a cause of infection [9,10]. Clinicians often perceive products that come in closed containers as sterile and use them as such even when they are not so marketed. Most would find surprising the frequency of contamination of oral medications (especially those of animal origin), ointments, nasal sprays, lotions, and mouthcare products [11].
The fomites discussed in this chapter include many entities for which no infection-causing potential has been established. Questions concerning the proper management of these items frequently confront infection control personnel. Even those skeptical of the infection-causing potential of environmental surfaces or ordinary objects used in patient care advocate processing and cleaning methods that may have a considerable financial impact. At times, procedures rest on aesthetic considerations. Carpets may not usually constitute an infection hazard, but fecal stains on them are unacceptable. Healthcare professionals may be convinced that microorganisms on the walls and floors play no role in causing human disease, but the lay public's perception is exactly the contrary. In an era of increasing attention to marketing and patient safety, visible dirt is undesirable. When standards have been set, formally or informally, they often have been based on the recognition of what reductions in microbial content can be consistently achieved with moderate resource use rather than on what levels prevent healthcare-associated infections (HAIs).
The fact that there are HAI outbreaks stemming from contaminated inanimate objects often is invoked as a basis for concern about endemic HAIs attributable to the inanimate environment. Publication bias brings forth the atypical, however. A single outbreak does not provide a basis for concern about environmental contamination.
Environmentally Altered Microorganisms
It seems paradoxical that environmental objects frequently can be contaminated by human pathogens but only rarely contribute to human infection. A potential explanation lies in the concept of environmentally damaged organisms. This concept has been rigorously demonstrated for Streptococcus pyogenes by Perry et al. and Rammelkamp et al. [12,13,14]. In a classic series of experiments, they studied streptococcal transmission in army barracks. Air, dust, or personal effects, such as blankets, were contaminated more often by streptococci when recruits had streptococcal illness or pharyngeal colonization. However, recruits who had been issued freshly laundered, Streptococcus-free blankets acquired streptococcal infection or pharyngeal colonization just as often as barracks mates issued highly contaminated blankets [12]. The authors assessed the infectious potential of naturally contaminated barracks dust. Dust samples, dispersed in small enclosures, produced between 3,500 and 56,600 streptococci/m3 air but did not result in pharyngeal colonization or infection in volunteers within the enclosures. Six volunteers had 17 direct inoculations of dust containing 1,800 to 42,000 streptococci onto the posterior pharynx. These resulted only in transient colonization lasting ≤30 minutes [12].
In contrast, fresh oropharyngeal secretions mixed with sterile dust and dried 4–8 hours produced streptococcal pharyngitis in two of eight volunteers after inoculation. Two of the remaining six volunteers developed pharyngitis on subsequent inoculation of smaller numbers of streptococci directly transferred on swabs of nasopharyngeal sections [14]. The designation of this phenomenon as environmental damage reflects a rather anthropocentric perspective. Streptococci presumably shift their metabolism to meet their needs. The physiologic basis of bacterial adaptation to desiccation has been explored [15] and involves substantial changes in internal constituents.
The loss of human pathogenicity associated with adaptation to the environment has been established for desiccated S. pyogenes. It seems plausible that other species have different adaptations to various environmental situations that also would reduce their pathogenicity for humans. For some organisms, however, it is reasonable to speculate that the required adaptations are less debilitating. “Water bacteria” in moist environments might be in metabolic states less different from their most pathogenic state. In contrast, Legionella spp. in their inanimate reservoir are hazardous, but person-to-person transmission does not take place. Viruses either are viable or not. Clostridium difficile as an environmental spore is more durable; if the spore is an infectious form for humans, environmental damage would not be relevant. These matters are, at present, largely unexplored.
Epistemology
With the foregoing generalities in mind, we confront the methodologic inadequacies that characterize most published studies asserting a causative role of inanimate objects in human disease. Evidence suggesting that a fomite has a role in causing disease due to a particular pathogen can be divided into seven categories. They are ordered here by the rigor with which they establish the point:
P.277
virus can survive when inoculated on hot-tub seats [16]. The finding was published despite the cogent words of the accompanying editorial [17], which pointed out the inconclusiveness of the observation.
Only evidence from categories 5, 6, and 7 should be considered strong enough to securely implicate a fomite.
Strain analysis and molecular markers of clonality, so helpful in many areas of epidemiology [22], are less useful in this epistemology (Chapters 10 and 15). Strain differences can negate evidence in category 2. However, strain analysis usually does not substitute for evidence in categories 5, 6, or 7. For instance, most hospital outbreaks of vancomycin-resistantenterococcus (VRE) or C. difficile have been due to a particular strain. Strain homogeneity, compared with extra-hospital isolates, can establish that the “cases” among patients are nosocomial in origin. Finding the outbreak strain in the environment does not, however, establish its causal role. The transmissions could result from hand carriage or some other person-to-person contact and the infected patients could be contaminating the environment. Only when the direction is clear (there is no likely way for patients to contaminate a hot water tank with Legionella spp., for instance) can establishing strain identity between patient's and fomite isolates establish environmental causality. There is one last distinctive aspect of the application of strain identification techniques to environmentally caused outbreaks. A bona fide outbreak may arise from heterogeneous strains, because environmental isolates are generally heterogeneous. For instance, nosocomial Aspergillus spp. isolates in an outbreak due to
P.278
a defect in air filtration will reflect the heterogeneity of environmental isolates [23].
Air
There has probably been concern about air as a vehicle for transmission of infection for as long as there has been recognition of the transmissibility of disease. Surely, more data have been generated for this mode of transmission than for any other. Entire books have been devoted to the subject [24,25], suggesting that concise summarization is difficult. For much of human history, air was considered the primary carrier of contagion. The word “malaria” reflects the recognition that proximity to swamps increased the risk of illness (the ancients were correct, in a sense, since mosquitoes fly). However, the advent of scientific medicine caused the pendulum to swing too far in the opposite direction. By 1910, no illness was any longer thought to be airborne. Large droplets travel up to 1–2 m through the air but are not borne on it. In the context of infection transmission, airborne means “borne on the air” rather than “transported through the air.” Large-droplet transmission is instead considered a type of contact spread [26] (Chapter 41). Only at the time of World War II did a more balanced view become established. Much of the subsequent work on air as a vector was done in the context of studies of biological warfare.
Although there are many concerns about air as a means of HAI pathogen transmission [27], the trend has been away from concern until very recently. Six articles dealing with airborne spread of HAI pathogens were presented at the initial International Conference on Nosocomial Infections in 1970, and Brachman estimated in 1970 [28] that between 10%–20% of endemic HAIs resulted from the airborne route. At the second International Conference on Nosocomial Infections in 1982, however, there was only one presentation about airborne organisms [29], and it dealt with the operating room. At the third and fourth International Conferences in 1990 and 2000, the topic was not discussed except for measles virus transmission.
Air may be sampled volumetrically or on settling plates (Chapter 10). Volumetric sampling produces quantitative data that are more readily conceptualized and seems more relevant to situations in which a pathogen is inhaled. It is the standard for rigorous studies. However, settling plates may be more appropriate to infections that result from settling organisms (e.g., wound infections), and such sampling can be performed without special equipment or expertise [30]. Unfortunately, there are few side-by-side surveys of air using volumetric and settling methods. In one study of air fungal content, 15-minute pairs compared volumetric sampling of 0.42 m3 air with settling on a 100-mm plate (79 cm2) [31]. A total of 127 sample pairs yielded total recoveries of 12,900 colony forming units (cfu) and 1,031 cfu, respectively, establishing one advantage for volumetric sampling. Furthermore, of the organisms captured by volumetric sampling, 9.2% were Aspergillus spp. while only 5.5% of the settling organisms were in that relatively buoyant genus, pointing out one of the complexities of interpreting settle plate results.
Assessment of the organism content of air is difficult because the concentration of certain organisms in the air is small compared with the volumes that can be conveniently assessed, producing considerable sampling error. Furthermore, there is tremendous variation in the microbial content of air, depending on location in the hospital, ventilation system, concurrent human activity, and proximity to sources of organisms. For a review of available air sampling methods, see Chapter 10 and Burge [32].
Few broad surveys of hospital air have complete microbial identification. In one set of studies, Greene et al. [33,34] found a mean organism count of 350–700 organisms/m3. The highest counts were in laundry-handling areas followed closely by other storage and disposal areas. The lowest counts were in operating and delivery rooms. Approximately one-third of the organisms recovered were gram-positive cocci, another one-third were gram-positive bacilli, and the remainder were gram-negative bacilli or fungi. Gram-positive cocci constituted a higher proportion of the organisms in operating rooms, gram-positive bacilli (presumably mostly Bacillus spp.) made up a higher proportion of the organisms in the laundry and waste storage areas, and gram-negative bacilli were found in relatively high numbers in corridors.
More detailed consideration of air as a means of transmission is best made by considering the specific organism. There is convincing evidence for airborne transmission for only a small number of pathogens (e.g., varicella-zoster virus, influenza [including Avian Influenza], measles, or Mycobacterium tuberculosi). For others, airborne transmission has been reported in rare situations, but most transmission is by droplet or contact transmission (e.g., Severe Acute Respiratory Syndrome Coronavirus (SARs-CoV), smallpox, Brucella spp., Pseudomonas pseudomallei, Coxiella burnetii, Chlamydia psittaci, Francisella tularensis, Bacillus anthracis, Legionella spp., Yersinia pestis, Pneumocystis carinii, Aspergillus spp., and other filamentous fungi). Additional organisms for example, mumps, rubella, Mycobacterium avium complex, S. aureus in very unusual circumstances, probably achieve airborne transmission. Some have argued that true airborne transmission (i.e., >3 feet) should be divided into intrinsic or opportunistic airborne-transmitted pathogens [27]. For most pathogens, humans (not the environment) are the major source of contagion. Control of airborne transmission in hospitals consists mainly of promptly identifying infectious patients and placing them in isolation rooms. Influenza control during community outbreaks is more complicated because of the high prevalence of infectious patients and personnel and because infectious persons may have subtle or no respiratory symptoms [35]. Preventing nosocomial influenza during
P.279
community influenza outbreaks is one of the few solid rationales for human traffic control within hospitals.
Staphylococcus aureus
A solid theoretic basis exists for concern about the importance of airborne transmission of S. aureus (Chapter 40). Noble [36], probably the most avid student of this matter, summarizes information bearing on the origin of airborne S. aureus as follows. Humans liberate ~3 × 108 squamae per day. Because the size distribution of airborne particles containing S. aureus (~4 to 25 µm in diameter] is nearly that of squamae and well above the diameter of naked, single S. aureus cells (~1 µm in diameter), it is presumed that most or all airborne S. aureus organisms are carried on these skin flakes. Because particles of this size become impacted on the nasal turbinates, a closed loop may exist: proliferation of S. aureus on the nasal mucosa, hand transfer of S. aureus to the skin, liberation on squamae, airborne transport of squamae, and impaction on the nasal mucosa. Hospital air contains ~0.7 S. aureus particles/m3 [37].
Outbreaks of S. aureus (and S. pyogenes) surgical site infections (SSIs) have been solidly linked to airborne spread from dispersers in the operating room (Chapters 35, 40). In this context, surgical gowns make direct contact improbable and masks make droplet spread improbable. However, the importance of air as a medium for endemic S. aureus transmission in other settings is less clear. The strongest positive evidence is that of Mortimer et al. [20], who studied acquisition of staphylococcal colonization in newborn infants housed in a special nursery that also was used for the care of known colonized infants. Extraordinary measures were undertaken to eliminate contact transfer of S. aureus from the index babies to the study babies. Nevertheless, at least 9/158 newborns became colonized. The authors offered as evidence the following points that these acquisitions were airborne: (1) Contact transmission did not occur; (2) index strains of S. aureus were recovered on settling plates throughout the nursery; (3) the infants were ≥2 m apart, making large-droplet transmission unlikely; and (4) in the study, infants tended to be colonized in the nose first, whereas in previous studies infants acquiring S. aureus by physical transfer tended to be colonized at the umbilicus first.
Wenzel et al. [38] have critically analyzed nine additional articles published between 1966 and 1976 that purport to establish airborne transmission of S. aureus. None of these additional studies provides even strongly suggestive evidence of airborne spread. A bit of negative evidence with respect to endemic operating room acquisition of S. aureus came from the National Academy of Sciences–National Research Council study [39] of the influence of ultraviolet radiation on postoperative SSI (Chapter 33). In the study, high-intensity ultraviolet light in the operating room reduced airborne bacterial counts, as measured on settling plates, by 52% or 63%, depending on the ultraviolet intensity used. At neither intensity was there a similar reduction in postoperative SSI rates. More recently, Bischoff et al. have documented that those colonized with S. aureus are more likely to disseminate the pathogen when they have an upper respiratory viral infection, the so-called Cloud HCW [40].
Gram-Negative Bacilli
Volumetric sampling of ordinary hospital air with recovery of pathogenic Enterobacteriaceae and nonfermenters has been rare. Available studies tend to focus on specialized areas of the hospital (especially operating rooms), use settling-plate methods, assess outbreak situations, or provide incomplete microbiologic identification. Klebsiella or Pseudomonas spp. [41], and other gram-negative organisms can be recovered from hospital air, but the best correlation with acquisition by patients is via handborne rather than airborne organisms [42].
Clinicians have been particularly concerned about the spread of Pseudomonas aeruginosa and Burkholderia cepacia from or to hospitalized patients with cystic fibrosis. Molecular strain identification techniques establish that institutional person-to-person transmission occurs [43,44], but they do not identify the mechanism. P. aeruginosa has been recovered on settling plates near patients with cystic fibrosis [45]. Institution of isolation protocols has been associated with reduced transmission rates. However, the potential interactions between patients with cystic fibrosis are many, and air is not established as a path of transmission.
Two studies have described an association between airborne gram-negative bacilli and endemic HAI. The first and more convincing situation resulted from a very unusual circumstance. The newly constructed Hines Veterans Administration Hospital had a novel chute hydropulping waste disposal system that introduced malodorous bacteria-laden air throughout the hospital. Air sampling near the system demonstrated >5,600 cfu/m3 of Pseudomonas organisms and Enterobacteriaceae (unfortunately, the relative amounts were unspecified) [46]. Concurrent continuous HAI surveillance found that the nosocomial BSI rate approximately doubled coincident with moving to the new hospital and fell to the baseline level after the chute hydropulping system was closed.
In a second study, carried out over a 5-year interval by Kelsen et al. [47], a significant positive correlation existed between the monthly rate of nosocomial respiratory tract infection in patients hospitalized in an intensive care unit and the average bacterial content of the air. During periods of heavy air contamination, the authors found an unusually high concentration of airborne gram-negative bacilli, ranging up to a P. aerugnosa content of 1,050 cfu/m3 and a Klebsiella spp. content of 315 cfu/m3. As the authors point out, the association may not imply that airborne
P.280
gram-negative rods directly cause nosocomial respiratory tract disease; it is possible that airborne bacteria seeded nebulizers [48] or another intermediate reservoir. The association may have resulted from a third factor affecting both bacterial content of the air and the HAI rate. Finally, it is possible that the air may reflect patients' illnesses rather than the reverse.
Ultimately, the best evidence that endemic HAIs do not often result from airborne gram-negative bacilli probably arises from the repeated failure to recover such organisms in air cultures obtained during outbreak investigations. Although these situations may be atypical, as negative evidence they are convincing because this may be expected to be the situation most likely to produce positive air cultures. The widespread belief that gram-negative bacilli do not survive for prolonged periods when airborne, if true, may provide additional evidence. Here the experimental support is more tenuous than one might wish. Under certain conditions of humidity, temperature, and physiologic state, Escherichia colican sustain up to 100% survival for half hours in microaerosols [49,50]. In general, E. coli survives better when aerosols are generated using broth cultures of organisms in relatively inactive states.
In summary, although there is insufficient evidence that airborne gram-negative bacilli constitute a source of endemic HAI to warrant changes in our current practices, situations that lead to high airborne concentrations of gram-negative bacilli should be avoided. Such situations include the use of aerosol-generating room humidifiers, which have been shown to cause considerable dissemination of Pseudomonas [51] or Acinetobacter spp. [52].
Legionella spp.
The original Legionnaires' outbreak included “Broad Street pneumonia” found among persons who remained outside the implicated hotel. These persons are presumed to establish airborne Legionella pneumophilia transmission, but they were not definitely proved to be due to L. pneumophilia. Subsequent outbreaks have been attributed to cooling tower contamination. Few of these studies have included long-term follow-up after cooling tower decontamination, nor have they demonstrated an association through strain identification. Although airborne transmission may take place, the emphasis for nosocomial legionellosis should be on control of contamination in the potable water supply.
Aspergillus spp.
Several lines of evidence strongly suggest that airborne Aspergillus fumigatus spores cause aspergillosis in immunosuppressed patients [53]:
It is likely that many other fungal organisms, such as Mucor, Fusarium, or Pseudoallescheria spp., also can in unusual situations be transmitted to patients through the air.
Occasionally, water reaches organic material within the hospital by penetration of the building shell, plumbing leakage, or condensation on chilled water lines or the inner surfaces of cold outside walls. Fungal growth may result within several days. Hospitals must act vigorously to stop the water leakage and promptly achieve drying. If growth of hazardous fungi has occurred, careful remediation is required.
Notwithstanding the occasional episode of in-hospital fungal growth, the majority of spores in hospital air are derived from outdoor air. Spores gain entrance into the hospital because of incomplete filtration—infiltrating around improperly seated filters, around window casings (especially when there is perpendicular wind), through entrances or via loading docks, and on the clothes of personnel and visitors. The better the filtration (i.e., the lower the spore counts), the more important the other sources of spores become. Spores may settle over time and be reaerosolized during manipulations of the HVAC systems while cleaning or during renovation that produces “mini-bursts” of spores. Explosive demolition produces very high Aspergillus spp. counts. Excavation can do so also, and renovation may [55] or may not [56].
Special efforts are required to protect bone marrow transplant (BMT) recipients (discussed later). Other patients are also highly susceptible to aspergillosis, including lymphoma and leukemia patients (especially those who are neutropenic for >7 days), solid organ transplant recipients, solid tumor chemotherapy patients, other recipients of cytotoxic therapy, patients with the acquired immunodeficiency syndrome, or steroid-treated patients. While special units for BMT patients are warranted, the potential for
P.281
competition in room assignment and issues concerning where one draws the line for the remaining groups of patients make it desirable to achieve low spore counts throughout the hospital. In BMT units counts of <0.02 pathogenic Aspergillus spp./m3 and, in the remainder of the hospital, counts of <0.05/m3 can be achieved without undue burden.
Human Immunodeficiency Virus
There has been anxiety about airborne human immunodeficiency virus (HIV) transmission, especially in the operating room and the autopsy department (Chapter 42). This anxiety persists partly because of the potential for aerosol generation by mechanical saws [57] or other activities. Clearly, splatter of blood and tissue happens in these locations, and scrupulous attention should be paid to barriers for the surgeon or prosector. The available studies do not, unfortunately, distinguish adequately between those droplets and tissue fragments that travel <2 m in smooth arcs to the ground and aerosols (droplet nuclei) that remain airborne and threaten persons at a greater distance from the aerosol generation point. The hepatitis B precedent would suggest that only operating room or autopsy personnel in direct contact with blood are at risk. Special devices to protect against airborne HIV transmission are not warranted.
Laser Plumes
Surgical lasers generate visible smoke that could, in theory, contain viable airborne organisms [58]. Although pathogen transmission has not been verified, safety guidelines have been published [59].
Water
Potable Water
Achievement of potability of water is a major public health activity, the discussion of which is beyond the scope of this book. Standard works may be consulted for details of water treatment and examination [60]. Verification of ordinary potability is of importance to infection control personnel only in hospitals with private water supplies where verification of water quality is a hospital's responsibility. It should be noted that U.S. federal drinking water regulations call for only one microbiologic assessment [61]: a coliform count (acceptable levels depend on sampling frequency but must average <1 per 100 ml). Even for community water, this sole criterion is probably inadequate, considering the variety of water sources, potential contaminants, and uses of water [62]. The European Community standards also include a limit on the total viable count [63]. Except for legionellosis (discussed later), there are few reports of HAI arising from drinking water [64]; thus, further consideration of water in this chapter focuses on specialized uses in the hospital.
Potable water supply systems must be protected by vacuum breakers or other devices to keep water from being sucked back into the system during unusual events. To save expense, some building designers plan separate potable and nonpotable water systems. These systems must never be interconnected. Common sense requires that the potable water system be used for all hand washing, bathing of patients, cooking, washing of food and utensils for cooking and eating, food preparation or processing, and laundry. Given the few valid uses for nonpotable water in the hospital and the difficulty in forever preventing cross-connection, the value of designing hospitals with separate nonpotable water systems is questionable.
Dialysis Water
Detailed standards for hemodialysis water have been prepared by the Association for the Advancement of Medical Instrumentation (AAMI) and accepted by the American National Standards Institution [65]. The standard specifies that water used to prepare dialysate shall have a total microbial count of <200 per ml and the dialysate shall have <2, 000 microbes per ml (Chapter 23). The rationale for the standard lies in studies carried out in the 1970s that indicated that pyrogenic reactions did not occur when dialysate had <2, 000 organisms per ml [66,67]. Bacteria do not cross an intact dialysis membrane, but endotoxin may. The viable bacterial concentration is a rough measure of the endotoxin concentration. The rationale for the stricter (<200 organisms per ml) standard for water used to prepare dialysate is that organism multiplication may take place within the dialyzer. This is a more important problem for recirculating systems than for single-pass systems [68], a distinction not recognized in the AAMI standard. In recirculating systems, dialyzed materials can provide nutrition to contaminating bacteria. Many types of water treatment devices are available for use in preparing dialysate [69]. A more detailed discussion is available in a Food and Drug Administration (FDA) technical report [70].
Hydrotherapy Pools and Tanks
A number of features of hydrotherapy tanks have generated concern that they may transmit infection. Patients using them may have active infection, which may introduce hazardous bacteria and organic debris; patients may be incontinent of feces; warm temperature, water agitation, and a high number of successive patients per unit volume of water reduce available chlorine; the internal channels of agitators are difficult to disinfect; and highly contaminated water may be brought into close contact with potential portals of entry, such as pressure sores, Foley catheters, and percutaneous devices (see Chapter 45). One outbreak of
P.282
Indeed, in a burn center where the infection potential of hydrotherapy tanks may be most severe, Mayhall et al. [73] have reported an Enterobacter cloacae BSI outbreak, which may have been associated with hydrotherapy transmission. To the extent that hydrotherapy tanks are similar to hot tubs and whirlpool spas, there is a more ominous possibility. Contamination of these water sources has resulted in P. aeruginosa folliculitis [74], urinary tract infections [75], and even pneumonia [76]. The danger from in-hospital hydrotherapy tanks is probably mitigated by higher standards of disinfection. In the outbreaks of community-acquired P. aeruginosa skin infections, even rudimentary standards of water maintenance were not in effect [74].
The Centers for Disease Control and Prevention (CDC) published recommendations for disinfection of hydrotherapy pools and tanks in 1974 [77]. For immersion tanks, the CDC recommended maintaining a free chlorine residual of 15 mg/L with a pH of 7.2 to 7.6, draining tanks between each patient's use, scrubbing out the tank with a germicidal detergent, and circulating chlorine solution through the agitator of the tank for ≥15 minutes at the end of each treatment day. For hydrotherapy pools, the CDC favored continuous filtration and the maintenance of free chlorine residuals of 0.4 to 0.6 mg/L. In the absence of continuous filtration, the CDC recommended potassium iodide and chloramine.
High-Purity Water
Distillation apparatus, reverse osmosis devices, and ion-exchange resin beds are all subject to contamination. Some hospital personnel erroneously assume this type of water is sterile. Distilled water, even if subsequently sterilized, may contain endotoxin. Febrile reactions caused by exposure to items rinsed in endotoxin-containing distilled water have occurred [78].
Water Bacteria
So-called water bacteria are organisms that proliferate in relatively pure water. The most adept species is B. cepacia. Carson et al. [79] reported on B. cepacia strains that could multiply to the levels of 107 / ml and remain at these high levels for weeks in distilled water of very high resistivity. P. aeruginosa follows closely behind in this ability [80]. Furthermore, P. aeruginosa strains adapted to distilled water are relatively resistant to disinfectants [81]. Acinetobacter calcoaceticus, an emerging pathogen, seems particularly well adapted to highly aerated water sources. An enrichment technique for isolation of Acinetobacter spp. from environmental samples using vigorous aeration has been described [82]. This feature of Acinetobacter spp. presumably accounts for its increased relative frequency of citation as a cause of humidifier or other respiratory device contamination [83,84]. Other water bacteria include Flavobacterium meningosepticum [85], other Pseudomonas species, Acromobacter species, Aeromonas hydrophila, Flavimonas [86], and certain nontuberculous mycobactria [87]. These last-named organisms are also relatively resistant to various disinfectants [88], including formaldehyde [89]. Among the water bacteria, P. aeruginosa and Acinetobacter [90] are unusual in that they also are common colonizers of healthy humans. Unprotected wet areas in a hospital should be considered contaminated with one or more water bacteria. These sources include tap water, drains and sinks, water baths, shower heads, flower water, ice machines, and water carafes.
Legionella spp.
Among the important HAI pathogens, Legionellaceae are the agents for which environmental sources are the most securely established. Person-to-person transmission of L. pneumophila is either very rare or nonexistent [91]. Nosocomial L. pneumophila pneumonia has been strongly associated with hot-water distribution systems and, perhaps, cooling towers (Chapter 31). Legionella micdadei appears to have a similar epidemiology [92]. Outbreaks of Legionella dumofi SSIs due to tap water contamination of fresh wounds appear to be a more atypical problem [93].
Muder et al. [94] have critically reviewed the mechanisms of transmission of L. pneumophila. Most of the initial outbreak reports, particularly outbreaks occurring in non-hospital settings, were associated with adjacent excavation or contaminated air-handling-system cooling towers. However, more recent hospital outbreaks have been securely linked to contamination of hot water systems. At the Wadsworth Veterans Administration Hospital in Los Angeles, where a large outbreak of nosocomial legionellosis took place over a period of several years, improvements in the air-handling system preceded efforts to eliminate Legionella spp. from the water system. Only the latter endeavor was followed by a reduction in the number of infections [95]. Many additional reports have attributed cessation of Legionella HAIs to reductions in the presence in hot-water systems of L. pneumophila. Unfortunately, with few exceptions [96], follow-up has been of more than one year's duration, and case ascertainment is unsure.
The way in which L. pneumophila contamination of hot water systems produces nosocomial pneumonia is not established. Presumably, inhalation of freshly aerosolized droplets predominates, although inhalation of particles (droplet nuclei) airborne from distant sources, aspiration of colonizing pharyngeal organisms, ingestion of drinking water, and contaminated respiratory therapy devices all remain possibilities [97]. An association of episodes with
P.283
the use of contaminated shower heads has been found in some [98,99], but not the majority, of investigations.
The need to keep hospital hot water systems free of Legionellaceae and the role of routine culture confirmation of the presence of organisms are unresolved. Until recently, CDC personnel opposed culturing hospital water not associated with episodes of infection [103]. In 1995, an ambivalent stance was adopted [104]. Nosocomial Legionella spp. pneumonia often goes unrecognized unless the possibility is avidly investigated [97], making it difficult for any hospital to be complacent about its Legionella situation. There is at least one cited episode of a hospital with a contaminated potable water system and fairly secure evidence of the absence of nosocomial legionellosis [105], but such reports are rare. Yu [106] has advocated a 1-year quarterly cycle of hospital water cultures in hospitals where identification of nosocomial Legionella pneumonia patients is uncertain. The availability of commercial media for Legionella cultivation makes such surveillance relatively simple although some expertise is required for confirmation of recovered isolates. Vigorously swabbing scale from shower heads is more sensitive than culture of flushed water, but detailed protocols for routine culture programs are lacking. Shower head scale and water from the base of hot water storage tanks should be included in surveillance. Hospitals recovering L. pneumophila and, perhaps L. micdadei, should strengthen clinical case-ascertainment methods.
Hospitals with nosocomial Legionella spp. pneumonia episodes and contaminated hot water systems must strive to eliminate the latter. Superheating the water (to as high as 77°C [170°F]) provides an immediate solution but may cause scalding of patients or HCWs. Instituting long-term solutions is more difficult. Persistent colonization of hospital potable water systems by L. pneumophila arises, at least in part, because the organism can tolerate low levels of chlorine for relatively long periods of time [107]. Hyperchlorination damages some plumbing system components, and raising the pH aggravates scale formation [108]. Other possible ways to eliminate colonization include use of instantaneous steam water heaters, ultraviolet light, chloramines, and ozonation [101].
Eyewash Stations
Clinical laboratories have eyewash stations for emergency eye flushing. These often go unused for months. There have been reports that water in these stations becomes contaminated with Acanthamoeba and other amoebas [109] capable of causing chronic destructive keratitis. Although no such infections have been reported, a weekly flush reduces the contamination.
Walls, Floors, and Other Smooth Surfaces
Maki et al. [18] performed a landmark study assessing the relationship between organisms on environmental surfaces and HAI. During 1979, the University of Wisconsin Hospital moved to a new facility. There was no change in the HAI rate at any patient-care site or due to any pathogen associated with this change. Cultures of floors, walls, or other surfaces (including air, water, faucets, and sink drains) showed very similar organism profiles in the old facility and, after 6–12 months of occupancy, in the new facility. In contrast, corresponding cultures taken in the new facility before occupancy were relatively devoid of common HAI pathogens. The constancy of HAI rates provides strong evidence that the association between hospital environmental organism content and HAI arises because patients contaminate the environment, not vice versa. It is important to realize some limitations of the Maki study. The two pathogens for which environmental content is of primary importance (Aspergillus and Legionella spp.) were not assessed, and the environmental cultures were not processed for anaerobes (e.g., C. difficile) or viruses.
This study virtually rules out the environment as an important vector for the assessed organism-object combinations and severely undercuts the rationale for concern about other combinations in the absence of specific data to the contrary. In fact, one is forced to question seriously even such relatively modest recommendations as the use of antimicrobial detergents in hospital cleaning, terminal disinfection of isolation rooms, special cleaning of objects removed from isolation rooms, and wearing gowns and gloves when entering the room of patients in isolation when no contact with patients is anticipated (except for vancomycin-resistant enterococcus or C. difficile).
Respiratory Syncytial Virus
One pathogen clearly transmissible in the hospital by fomites is respiratory syncytial virus (RSV) (Chapter 41). Indirect evidence suggests that RSV transmission happens through contact; inoculation of RSV onto nasal or eye membranes causes infection quite efficiently [110]. Moreover, this virus survives for several hours on smooth
P.284
surfaces [111]. Direct evidence of fomite transmission is now available [112]. Volunteers who entered a hospital room that had just been vacated by an RSV–infected patient became infected with RSV >50% after handling objects in the room and touching their eyes and nose as often as volunteers who cuddled RSV–infected babies. Volunteers sitting in the same room with an RSV–infected baby did not contract the illness. The relative importance of hand and fomite transmission is unknown, but it is of interest that RSV survives approximately 10 times better on smooth surfaces than on skin [111].
Clostridium difficile
Strain analysis techniques have unequivocally established that in-hospital transmission of C. difficile can take place (Chapter 33) [113]. Variation in hospital C. difficile transmission would provide a satisfactory explanation for the apparently wide variation in rates of C. difficile colitis in different hospitals. As with many HAI pathogens, increased environmental presence of C. difficile is associated with infected patients. In their excellent review, McFarland and Stamm [114] found five supportive studies. With regard to environmental concerns, what distinguishes C. difficile is its ability to form spores with the consequent prolonged survival of the organism in the environment and the plausibility that the spores retain full infectiousness. However, a controlled study of glove use suggested that most C. difficile transmissions arise from carriage by hands [115]. The same group has used restriction endonuclease strain analysis to establish that C. difficile acquisition does not geographically cluster within wards and is not more likely to be transmitted to a subsequent bed occupant [116]. No special environmental cleaning techniques for C. difficile contamination have been formally advocated, although the use of an agent such as bleach that is effective against spores is recommended. Emergence of a toxinotype III strain that is positive for binary toxin, an 18-base pair deletion in tcdC, and has increased resistance to fluoroquinolones is of concern.
Hepatitis B Virus
Concern about environmental hepatitis B virus (HBV) transmission arises from several lines of evidence. Clinical laboratories and hemodialysis units, areas frequently contaminated by blood, were foci of HBV transmission throughout the 1970s (Chapter 42). Many of the ward-acquired, and an even larger fraction of the laboratory-acquired, HBV episodes occurred without recognized percutaneous inoculation of blood. A decline in the incidence of hepatitis among HCWs began in the mid-1970s (before the introduction of HBV vaccine) when concern about blood contact became widespread [117]. Approximately 30%–40% of community-acquired episodes of HBV cannot be ascribed to sexual contact, needle sharing, or therapeutic blood component exposure [118]. Hepatitis B surface antigen (HBsAg) may be antigenically detected on surfaces in hospital areas likely to have been blood contaminated [19]. Surfaces not visibly contaminated with blood may also yield HBsAg. Even today, blood contamination can frequently be found on patient-care items [119]. In blood, HBV remains viable for ≤1 week after desiccation at room temperature [120], although inactivating it with disinfectant is not difficult [121]. Very high dilutions of HBV-containing blood can transmit hepatitis B.
These lines of evidence do not establish a role of the environment in HBV transmission. Coincident with efforts to eliminate or decontaminate environmental blood was the adoption of segregation of HBsAg-positive patients in dialysis and more widespread recognition of the hazard of needle sticks. Nevertheless, when contaminated objects have been in close proximity to a portal of entry, such as the finger platform of an automatic finger-stick device, HBV transmission by inanimate objects has been verified [122].
To clean blood spills, the CDC has recommended the use of any chemical germicide that is approved by the U.S. Environmental Protection Agency (EPA) as a “hospital disinfectant” and is tuberculocidal [123]. Because no contact time was specified, the recommendation is not that a tuberculocidal standard be met. Nevertheless, OHSA may arbitrarily enforce the germicide choice strictly. In areas where large blood spills are commonplace, such as around operating room tables, even this recommendation seems excessive. As long as such units have specialized cleaning protocols and good protective equipment, any hospital disinfectant should suffice. With large spills of cultured or concentrated agents in the laboratory, the contaminated area should be flooded with germicide before cleaning. Otherwise, the area should be cleaned and then decontaminated.
Viral Hemorrhagic Fever Agents
The Centers for Disease Control and Prevention (CDC) regards environmental surfaces to be a potential vector of viral hemorrhagic fever agent transmission. The 1980 recommendations [124] for the treatment of patients with Lassa fever and other acute viral hemorrhagic fevers included special isolation units with exhaust air filtration, use of a chemical toilet, disinfection of all items taken from the patient's room, and disinfection of the vacated room with gaseous formalin or paraldehyde. The CDC 1988 revision [125] restricted concern to patients with confirmed or suspected hemorrhagic fever due to the agents of Lassa, Marburg, Ebola or Crimean-Congo hemorrhagic fever. CDC statements in 1995 and 2005 [126] recommended the use of a negative pressure room with an anteroom and HEPA respirators for entering the room of patients with prominent cough, vomiting, diarrhea, or hemorrhage; minimizing laboratory testing; autoclaving, incinerating, or using bleach in washing linen; use of a chemical toilet or bleach disinfection of excretions and fluids
P.285
before discarding them in the sewer; incineration or decontamination of solid medical waste; and processing clinical specimens in a class II biological safety cabinet following biosafety level 3 practices.
Other Environmental Factors
Carpets
Carpeting a floor increases the microbiologic content per unit of floor surface by approximately four orders of magnitude [127]. Contaminating organisms include S. aureus, E. coli,and, more rarely, Pseudomonas. After removal of carpeting from hospital environments, the carpet content of S. aureus remains stable for more than a month and of other organisms for ≤6 months. In the best-controlled study, however, the total bacterial content of air was apparently unaffected by the presence or absence of carpet in the sampled area [128]. For areas with carpets, the air content also was not significantly influenced by vacuuming frequency (daily, every other day, or every third day) [128]. Unfortunately, these studies were carried out before the widespread adoption of large, self-propelled, high-pressure cleaner-extractors, which do create transient increases in air fungal content.
The infection hazard of carpets may be more important when patients have direct contact with carpeting (e.g., in pediatric areas) or when patients use wheelchairs. Wet machine cleaning has been associated with Aspergillus flavus proliferation. However, it has yet to be demonstrated that any HAI has arisen from a carpet. In recent years, manufacturers have marketed carpets with antimicrobial substances. As yet, these have not been rigorously assessed in independent studies.
Air-Fluidized Beds
Designed to prevent pressure sores by “flotation,” air-fluidized beds have features posing a potential for infection transmission. Flotation is accomplished by driving air up through a 25-cm-deep layer of silicon-coated, soda lime glass microspheres 50–150 µm in diameter. The microspheres are held in the bed by a monofilament polyester filter sheet with openings of ~37µm through which the microspheres cannot pass. Disinfection of the beds is accomplished by sieving out clumps of beads and organic debris and then operating the bed at high temperature and air flow to inactivate organisms by heat, abrasion [129], and desiccation.
Initial anxieties about the infection hazard from air-fluidized beds have largely dissipated. Beds spiked with Staphylococcus epidermidis, P. aeruginosa, or Bacillus subtilis did not cause airborne dissemination of these organisms, even shortly after inoculation [130]. The air over beds contaminated by use did not contain more organisms than control air over ordinary beds [131]. A single report of infection transmission, due to Enterococci, has appeared [132], suggesting that attention to proper decontamination of these beds is important. There remains the theoretic possibility that the air fluidization process renders airborne those organisms that usually remain harmlessly attached to surfaces (e.g., M. tuberculosis), and a study of the beds in the most heavily contaminated contexts has not been undertaken.
Soap
Given the emphasis on hand washing and hand hygiene, it is surprising that the problem of soap contamination is not better studied. Most soaps are not marketed as sterile. Recent outbreaks associated with either intrinsically or extrinsically contaminated soap reemphasize the potential danger of soaps as a cause of HAIs [133]. Other outbreaks have demonstrated the potential danger of intricately contaminated iodophors [10]. It is reasonable to postulate that it is advantageous to wash the hands with sterile soap (liquid or leaf) dispensed from forearm- or leg-operated dispensers that are resistant to contamination. The use of contaminated hand lotion has been implicated in a P. aeruginosa outbreak [9].
Data confirming the expected contamination of in-use bar soap have been widely disseminated in the promotion of dispensed liquid soap [134]. However, data comparing the microbial burden on hands washed using a contaminated soap bar with that using uncontaminated nonmedicated soap are unavailable. At the least, it seems prudent to reduce the microbial content of soaps by using disposable liquid soap containers, thoroughly cleaning reusable liquid soap containers, or, if bar soap is used, purchasing small bars and providing soap racks that permit water drainage. The relative merits of these alternatives await additional study. In countries with limited resources, care should be taken not to use bar soap and, not to “top off ” or refill liquid soap dispensers without disinfecting them between refilling to ensure that single-use paper towels are available for hand drying or to use waterless agents. The recent recommendations for use of waterless alcohol-based hand-hygiene agents enhance the availability of hand-hygiene facilities at the bedside, reduce the risk of soap contamination, and obviate the need for towels for drying (see Chapter 3).
Flowers
Flowers pose two theoretic infection hazards. Vase water inevitably contains large concentrations of potential HAI pathogens [135], and decaying organic matter in the dirt of potted plants provides a substrate for fungal growth. Although there are no convincing data establishing vase water as a seat of HAIs, many hospitals bar flowers in water and potted plants from the rooms of immunosuppressed or ICU patients. If vase water were to be disposed of
P.286
gently and patients, personnel, and visitors washed their hands after touching the water, little danger should arise. Unfortunately, achieving uniform compliance with these precautions is improbable.
Animals
Sanctioned animal contact with hospital patients is of several types: blind or disabled patients or personnel may be accompanied by seeing eye or other service animals, pets may be brought to patients, animals may be used to entertain patients or provide pet therapy, and research animals may be housed in areas near patient-care units. Although Q fever is the only zoonosis shown to have been associated with epidemics in healthcare facilities, >100 organisms infect both humans and other animals [136]. Knowledge of transmission mechanisms of these organisms suggests several prudent measures.
Persons handling seeing eye or other dogs should be sure that the dogs are vaccinated against rabies, appear to be free of ectoparasites, and are healthy (in particular, ringworm should not be present). Arrangements should be made for walking the dogs and assuming responsibility for disposal of animal excreta. Pregnant ewes used in research centers have been the source of outbreaks of Q fever [137]. Hospitalized patients have not been affected in these outbreaks; however, airborne transmission to personnel having no direct contact with pregnant ewes has occurred. Although the hazard appears to be most severe at or near parturition [138], it seems prudent to bar all contact between patients and pregnant ewes and to ensure that pregnant ewes used for research are never, even during transportation, in areas from which airborne spread to patients can take place. More stringent recommendations for protection of personnel working with pregnant ewes have been published [139].
Certain animal contacts with children seem inadvisable in any circumstance. Reptiles cannot be reliably certified to be free of salmonellosis, wild carnivores (e.g., skunks, ferrets, raccoons) and bats pose an unacceptably high risk of rabies, and birds of virtually any species may transmit C. psittaci. Any contact with animal urine should be followed by hand hygiene and, if appropriate, more extensive disinfection procedures because of the possibility of leptospirosis. Contact with mouse or hamster urine also is hazardous to the immunocompromised patient because of the possibility of transmission of lymphocytic choriomeningitis virus. Household pets also can contaminate the hands of HCWs, who then can introduce the pathogen. This was illustrated by an outbreak of Malassezia pachydermatis in neonatal ICU patients [140].
Linen
Considering how heavily contaminated soiled linen is, it is remarkable how rarely it causes infection. Laundry workers, who have prolonged close contact with soiled linen, seem to be at risk only as a result of exposure to blood-contaminated sharp implements, hepatitis A [141], or other enteric pathogens [142], or unusually infectious organisms, such as C. burnetii. None of these dangers represents a meaningful hazard to patients. It seems prudent to handle soiled linen gently to reduce the dispersal of microorganisms in areas dedicated to the care of patients. Beyond that, there is little basis for employing special procedures. Given the improbability that soiled laundry reposing in a partially filled hamper adds organisms to the environment (much less causes HAI), it is difficult to understand the emphasis that hospital inspection agencies have previously placed on closing soiled-linen hampers.
Clean laundry, even after cold water processing, contains few pathogenic organisms. Sheets have a total aerobic colony-forming unit count of ~0.2 cfu/cm3, and terry cloth items have ~2 cfu/cm3. The profile of contaminating organisms (Bacillus spp, 58%; coagulase-negative staphylococci, 25%; Corynebacterium spp, 18%) after cleaning is markedly different from the prewash profile. Pathogenic species are rarely found in washed linen. The proper handling of clean linen during transportation and storage is probably the most important determinant of the microbial content at the time of use. Meyer et al. [143] studied newborn ICU laundry that had been washed at 75°C (167°F), dried at 96°C (205°F), and carefully handled. Rodac contact plates showed no organisms one-third of the time and >10 colonies per contact plate only 9% of the time. Linen near the top of the stack had a higher incidence of positivity and a higher number of colonies per plate than linen in the middle of the stack, suggesting that handling was the source of transfer of organisms.
Nosocomial Bacillus cereus infection has been attributed to clean linen [144,145]. The reported outbreak consisted of B. cereus umbilical colonization without clinical signs of infection in normal neonates and neonates in a special care unit. The source was considered to be contaminated clean diapers because the implicated B. cereus type was found in washed diapers and in the laundry machine. Because the implicated B. cereus type also was recovered from the hands of nursing staff, this attribution is unconvincing. Other reports of HAI due to clean linen—tinea pedis in a nursing home [146], staphylococcal disease in newborns [147], or urinary tract infection [148]—are likewise not persuasive.
The final revision of the American Hospital Association's Infection Control in the Hospital [149] recommended autoclaving linen for patients “particularly susceptible to infections,” such as burn patients and in the nursery. The American Academy of Pediatrics [150] supported this recommendation until it softened its stance in 1983 [151]. No consensus body has such a recommendation at present. One line of argument against autoclaving linen arises after consideration of the panoply of techniques required to maintain sterility until the linen reaches the point of use. Applying these procedures to autoclaved linen would be
P.287
burdensome and costly. Only one rationale for autoclaving laundry seems plausible. Laundry dried in unfiltered air becomes contaminated by A. fumigatus, and in specialized patient-care units with a very low fungal spore content and low air change rates, a few introduced spores can contribute a substantial portion of the ambient spores [152].
Ultraclean Protective Environments
The ultimate expression of concern that environmental organisms pose an infection hazard is the ultraclean protective environment. When fully developed, these environments have HEPA air filtration with horizontal or vertical laminar airflow; sterile food or food with a low organism content; frequent disinfection of walls, floors, and other environmental surfaces; sterile linen and drinking water; toilet water disinfection; sterile booties, gowns, caps, and gloves for HCWs and visitors entering the room; and elaborate protective garb for patients leaving the room. Patients placed in such an environment are generally given oral nonabsorbable, topical, or systemic antimicrobials. This package of protective measures has been termed a total protective environment, life island, protected environment, or barrier isolation.
It has long been recognized that these special efforts can produce environmental surfaces and ambient air with markedly diminished organism content [153]. More important, a meta-analysis [154] of random allocation trials of various forms of ultraclean protective environments suggested that this package of techniques produces a statistically significant reduction in the incidence of infection. Of 10 trials [155,156,157,158,159,160,161,162,163,164], 5 showed a statistically significant reduction in overall, severe, or fatal infections. Of the remaining studies, three trials showed a trend to fewer infections in the protected patients, and one did not report infection rates. This infection prevention effect was generally noted after the second week, a finding in agreement with the view that there is a lag between becoming colonized with a HAI pathogen and subsequent infection.
However, the use of ultraprotective environments remains controversial for a number of reasons:
Clearly, special efforts are warranted to lower the airborne fungal spore concentration in the rooms of highly immunosuppressed patients [165]. It is important to consider separately several features of air purification systems. Top-of-the-line bag filters probably remove nearly all fungal spores, but many hospitals prefer HEPA-filters because they add relatively little capital expense and meet standards more directly related to microbial filtration. The CDC has recommended HEPA filtration [104]. It may be desirable to place duct insulation outside the ducts [166].
P.288
Placing filters at the point of entrance of air into a patient's room permits safe local maintenance while the room is otherwise unoccupied and accommodates malfunction or maintenance of the central system. Sealing the room and placing it at positive pressure reduce infiltration of spores from the outside or adjacent hospital areas. Increasing the air change rate minimizes the potential exposure of a patient to infiltrating or introduced spores [152].
Air change rates of ≥10 per hour are desirable for BMT patients. Rates above 15–25 per hour are best accommodated by placing fans and filters next to the rooms because the caliber of ducts conveying air to and from central systems would have to be too large. Laminar airflow (a misnomer because objects in a room cause considerable turbulence) units are best thought of as ultrahigh air change rates. Air change rates of 100–400 per hour can be achieved. Attention to in-room air flow patterns [167] may lead to improvements in the safety of patients without requiring extra resources. Placing patients with infectious diseases or infected with the most infectious agents (e.g., M. tuberculosis) toward the exterior ward wall with windows in hospitals in countries with limited resources may reduce the risk of such agent transmission. Being cognizant of the air flow patterns in the room also may facilitate patient placement to minimize the risk of pathogen transmission.
Concerns are associated with other efforts to eliminate a patient's exposure to environmental organisms. Food with low organism content is unpalatable. Organisms on surfaces that do not come in contact with the patient probably are harmless. Elimination of environmental organisms can be very difficult. In one ultraprotective unit, there was a prolonged struggle to eliminate an unusual Pseudomonas spp. from toilet bowl water. Notably, although the organism was present for 20 months, no instance of HAI or colonization due to the organism was identified [168].
Except for fungal spore control, ultraprotective environments are not yet an established infection control measure. Even their advocates do not believe they truly are indicatedexcept for patients undergoing BMT or intensive chemotherapy likely to produce >25 days of granulocytopenia [169]. What is critically needed is analysis of the relative benefit of the components of the protective package.
Disinfection and Sterilization
This topic is addressed in Chapter 20.
Definitions
Sterilization means the complete elimination or rendering nonviable of all microorganisms, including all spores (Table 19-1). Nonviable is best taken to mean the irreversible loss of the ability to propagate indefinitely [170]. Ultraviolet light, although lethal, does not interrupt germination and temporary growth. Conversely, organisms seemingly killed by mercury can be resurrected by compounds that displace mercury from sulfhydryl groups. Disinfection is divided into three levels [171]. High-level disinfection means the elimination of all viruses and vegetative microorganisms and most, but not necessarily all, bacterial or fungal spores. Intermediate disinfection means the elimination of all vegetative pathogenic bacteria, including M. tuberculosis, but not necessarily all viruses (nonenveloped or smaller viruses are more resistant to disinfection) or spores. Inactivation of M. tuberculosis is used in this definition not primarily because of concern about M. tuberculosis contamination. Mycobactericidal capacity is used because the organism is relatively resistant to disinfection compared with other vegetative bacteria, and a procedure to assess mycobactericidal activity has been established by the Association of Official Analytical Chemists (AOAC) [172], even its procedure has been challenged [173]. Low-level disinfection, roughly equivalent to sanitization, means the elimination of most pathogenic bacteria. Cleaningmeans the removal of all visible debris. All items should be scrupulously cleaned before disinfection because disinfecting methods may not penetrate debris. Antisepsis is the application of compounds to skin or mucous membranes to reduce microorganism content substantially.
Although the preceding definitions correspond best to practical use requirements, the Environmental Protection Agency (EPA), the main regulatory agency for disinfectants and sterilants until 1993, used a noncongruent classification of chemical germicides. Sporicides meet an AOAC standard for spore destruction [172]. They achieve sterilization or high-level disinfection depending on contact time. Hospital disinfectants inactivate Salmonella choleraesuis, S. aureus, and P. aeruginosa in highly specified AOAC tests [172]. Disinfectants andsanitizers meet other tests. The EPA registration categories made no reference to effectiveness against M. tuberculosis, the critical distinction between intermediate- and low-level disinfectants or to effectiveness in inactivating all viruses, the critical distinction between high- and intermediate-level disinfection. In the United States, the Food and Drug Administration (FDA) is now the regulatory agency for germicides that are sterilants or used for high-level disinfection. Its criteria for labeling are sufficiently stringent that few germicides have achieved sanction.
Kinetics of Microbial Killing
It is generally assumed, although not always supported by experimental evidence, that most microbial inactivation processes follow a “one-hit” killing curve. This presumption is equivalent to asserting that all the organisms in the population are equally susceptible to the process. These presumptions can be restated mathematically as follows. The number of microorganisms killed is proportional to the number present, and the proportion does not change as the population of remaining organisms decreases. When the logarithm of the concentration of organisms is displayed on the vertical scale and time on the horizontal scale, this
P.289
P.290
relationship results in the familiar straight-line killing curve. The slope of the killing curve is the measure of the rapidity of organism destruction. It often is expressed as the decimal reduction time, the time interval required to bring the concentration of organisms to 1/10 its previous concentration (i.e., 90% destruction). The difficulty in validating one-hit kinetics arises because of technical obstacles to experimentally ruling out the possibility that a very small fraction of the starting population of organisms is more resistant to killing. The potential difficulty in killing the last few (possibly more resistant) contaminating organisms is one basis for the overkill present in most sterility standards.
|
TABLE 19-1 |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
The preceding kinetics analysis establishes the importance of exposure time in accomplishing microbial destruction. A perfectly acceptable disinfection process will fail if not applied for sufficient time. If extremely high numbers of organisms must be inactivated with a very high probability that no survivors remain (e.g., manufacture of some vaccine), prolonged exposure times may be required. Furthermore, a given process may be sanitizing, disinfecting, or sterilizing, depending on the length of time it is applied.
Microbial Safety Index
The kinetics analysis also establishes that the operational assessment of sterility is a probabilistic assertion, not an all-or-nothing phenomenon [174]. This fact has led to the recommendation that the label sterile be supplemented by a microbial safety index (MSI) [175], defined as the absolute value of the logarithm of the probability that the item is contaminated.
For example, an item with an MSI of 3 would have a probability of 1 in 1,000 of containing a viable microorganism. As a practical matter, establishing that an item in a lot has an MSI>3 is extremely difficult by direct microbiologic assessment. With even the most rigorous culture technique, it is difficult to avoid introducing contamination at a level much less than 1 per 1,000 cultured items. Furthermore, the mathematics of sterility testing is unfavorable. For instance, to establish with 95% confidence that a lot containing 10,000 items is contaminated at a rate of <1 per 1,000, almost 3,000 of the items must be cultured and found sterile.
Administrative Issues
The FDA requires that reusable medical devices be sold with specific instructions regarding reprocessing methods [171]. The use of alternate methods may invalidate a warranty or create a medicolegal dilemma. The latter problem arises if a product failure damages a patient. The manufacturer may try to shift liability to the hospital because the product was not used according to instructions. Manufacturers may thus escape the stringency of strict liability for product failure.
These same considerations apply to reprocessing disposable items. Through the early 1980s, relevant standard-setting organizations lined up fairly solidly against reprocessing disposable items. The CDC recommended in 1982 that “no disposable object designed for sterile, single use should be re-sterilized” [176]. This restriction was rescinded in 1985 [177]. Through 1984, the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) flatly opposed any reprocessing. However, the 1995 JCAHO standards were less restrictive, merely requiring hospitals to have written policies that address reprocessing methods [178].
A 1977 FDA policy guide assigned full responsibility to the hospital when disposable medical devices were reused [179]. However, the FDA guide explicitly sanctioned the reuse of disposable items when the facility can establish that the item can be cleaned and sterilized adequately, that its “physical characteristics or quality are not adversely affected by their reprocessing,” and, somewhat redundantly, that the product remains safe and effective for its intended use. None of these statements addresses the resterilization of an unused item. Occasionally, an item is removed from its package or its package has been damaged, but the item has not been used. Consistency requires that these items also be resterilized.
P.291
On August 14, 2000, the FDA released the document “Enforcement Priorities for Single-Use Devices Reprocessed by Third Parties and Hospitals” to provide guidance to third-party and hospitals reprocessors about their responsibilities as manufacturers engaged in reprocessing devices labeled for single use under the Federal Food, Drug, and Cosmetic Act (the FFDC Act), as amended by the Safe Medical Devices Act of 1990, the Medical Device Amendments of 1992, and the Food and Drug Modernization Act of 1997. Third-party and hospital reprocessors of single-use devices (SUDs) are subject to all regulatory requirements currently applicable to original equipment manufacturers, including premarket submission requirements (Section 513 and 515 of the FFDC Act; 21 Code of Federal Regulations Parts 807 and 814) (www.fda.gov/cdrh/ohip/guidance/1333.pdf).
A key question, begged by all the aforementioned bodies, is “How does one determine whether an item is disposable?” At present, the manufacturer makes the determination. An item is disposable if it comes in a package labeled with the words disposable, single use only, or the like. Some manufacturers have added such language to packages of products previously marketed with resterilization instructions. Indeed, manufacturers have little incentive, at least in the short term, to do otherwise. Labeling an item as disposable minimizes liability and maximizes sales volume.
The most compelling case for reuse of disposable items has been made for dialyzers [180,181,182] (see Chapter 23). First use of hollow-fiber dialyzers more often may be associated with mechanical failure and systemic reactions (fever, chest pain, transient fall in white blood cell count) due to chemicals leaching out of the membrane or increased complement activation by new dialysis membranes. Some first-use-type reactions continue to occur, however, with reused dialyzers [183]. Other items may be very expensive and capable of withstanding reprocessing methods. Because resterilization of an item costs a hospital between $10–$20, depending on the time required to clean and package it and the sterilization method used, the impetus to reuse exists only for expensive items. A detailed protocol for reuse of specific items has been successfully employed [184].
The JCAHO requires written hospital policies regarding decontamination and sterilization activities, the performance of sterilizing equipment, and the shelf life of all stored sterile items [178]. The CDC guidelines recommend weekly biologic monitoring of all sterilizers [177]. When an implantable device is sterilized, a biologic indicator should be used and found sterile before the device is implanted. A chemical indicator should be visible on the outside of all sterilized packages. Careful follow-up of unconverted indicators should be undertaken because investigation often turns up significant problems [185].
Choice of Sterilization or Disinfection Level
Support continues for Spaulding's classification scheme indicating the level of sterilization or disinfection required for various items [171,177,186]. Critical items enter tissue or the vascular space. Semicritical items come into contact with mucous membranes or nonintact skin. Noncritical items touch intact skin. Critical items are generally held to require sterilization, semicritical items to require high-level disinfection, and noncritical items to require intermediate- or low-level disinfection [171,176,187]. Virtually all germicides are effective against HIV [188].
High-level disinfection is, in fact, rather difficult to achieve. Most germicides require 20–45 minutes to attain tuberculobactericidal activity, exposure times well in excess of common usage [189]. Furthermore, if a device really has to be at a state of high-level disinfection at the time of subsequent use on another patient, it would have to be subject to sterile water rinsing, manipulation using sterile technique, air drying with filtered air, and protective wrapping. Such precautions are rarely part of hospital practice [189], nor are they called for in many specialty societies' published guidelines for the reprocessing of semicritical items [190]. This amounts to an acknowledgment that organisms carried over from the previous patient are the primary target of reprocessing techniques. This is a rational emphasis provided that the disinfected device is protected from “water bacteria” by complete drying and from gross recontamination or hand contact before subsequent usage.
The assertion that all semicritical items must be processed by high-level disinfection is difficult to justify. To make this claim is to suggest that items coming into contact with mucous membranes can be contaminated with no more than a few bacterial or fungal spores. High-level disinfection seems unwarranted for items that will touch normally contaminated mucous membranes, such as the mouth or the colon. The distinction made between mouthpieces, for which some authorities have recommended high-level disinfection [189], and silverware is difficult to understand. For items in contact with the gut, elimination of carryover enteric pathogens is the goal. Unfortunately, assessment of the ability to inactivate small nonenveloped viruses, which are the most resistant enteric pathogens, is not routinely available.
The evolution of the category of intermediate-level disinfection is intriguing. It was not included in the CDC guidelines published through 1983 [191]. The category was defined in the 1985 CDC revision [177], although there were no specific recommendations for how to achieve it nor distinctions made between items requiring low- vs. intermediate-level disinfection. In 1987, Rutala and Weber's table combined low- and intermediate-level disinfection [192]. In the 1990 Association for Practitioners in Infection Control (APIC) [193] guideline, the categories were separated and specific indications given for each. Both were described with maximum exposure times, suggesting that the briefest contact with the disinfectant suffices. Unfortunately, although the difficulty in achieving and maintaining high-level disinfection establishes the usefulness of
P.292
a less-intensive disinfection level, there remains practically no rigorous study or body of evidence providing a foundation for the category of intermediate-level disinfection.
Special issues regarding certain devices should be recognized. Nebulizers produce, by design, particles that become deposited in the alveoli. Accordingly, nebulizer cups and solutions intended for nebulization must be sterile. Endoscopic retrograde cholangiopancreatography is potentially much more hazardous than all other forms of endoscopy as a result of the vulnerability of partially obstructed biliary tracts to infection and the severe nature of acute cholangitis. Contamination of the biliary tract can arise if the water channel in an endoscope used for the procedure holds contaminated water. When water is expelled from the catheter and then drawn back into the suction channel, organisms introduced into the suction channel can be picked up by the cannula before introduction above the ampulla of Vater.
Tonometers pose special difficulties. Numerous adenovirus outbreaks have resulted from inadequate disinfection procedures [194,195]. Adenoviruses, which are small and lipid free, are relatively difficult to disinfect. Furthermore, tonometer tips are expensive, harmed by many disinfectants, and used frequently. In addition, pneumotonometer tips have a cavity that can retain germicides with the potential for subsequent damage of a patient's cornea. The American Academy of Ophthalmology's recommendations for simple alcohol wiping do not achieve even intermediate-level disinfection [196]. Automated reprocessing machines also have produced disinfection failures [197]. Standards for evaluating these machines have been published [198].
Dental items are upgraded beyond the general scheme because of a consensus set of recommendations [199].
Many endoscopy systems, ultrasound probes, and other semicritical items are now marketed with disposable sheaths. In theory, the underlying item should need no reprocessing whatsoever after the sheath is discarded. Unfortunately, there is relatively little independent assessment of the integrity and durability of these sheaths. Clearly, if a defect in the sheath is detected after use or if the underlying item is visibly contaminated with a patient's secretions, the item should be reprocessed as if it had been used without a sheath. Otherwise, it is probably sufficient to reprocess the item using low-level disinfection.
Steam Sterilization
Steam sterilization is highly reliable and is the method of choice when the device can tolerate the procedure. Nevertheless, the subtleties to its use sometimes go unrecognized. Steam is more than an efficient conveyor of heat. The water molecules participate in the denaturation of proteins and the disruption of other complex molecules. Accordingly, it is essential that steam reach all the surfaces to be sterilized. In gravity displacement autoclaves, the introduced steam, which is less dense than air, forces air down and out through the autoclave drain. Devices with depressions that are not placed on their side or that have curved lumens will not be completely exposed to steam. The American Association of Medical Instrumentation (AAMI) standard for 132°C (270°F) sterilization assumes that this problem can be overcome by extending the cycle to 10 minutes [69]. Unfortunately, if steam does not reach the surface, this is equivalent to dry heat, a process that is generally held to require 2 hours of exposure. The penetration of steam into wrapped packages, porous materials, or in over-packed chambers also is not secure in gravity displacement autoclaves. These problems are mitigated in vacuum displacement autoclaves, which are evacuated before the introduction of steam. Pulsed vacuum autoclaves are even more efficient because they go through several cycles of vacuum and steam replacement, thus more reliably eliminating air.
Flash autoclaving also has attendant problems. The term itself is used variably to refer to short-duration, high-temperature steam autoclaving; the autoclaving of devices without wrapping; gravity displacement autoclaving; or some combination of the foregoing procedures. There is doubtlessly a need for rapid sterilization of low-inventory instruments that inadvertently become contaminated during surgery or highly tailored implantable items for which it is difficult to maintain a complete sterile inventory. Recent evidence suggests that the widely accepted 3-minute standard for 132°C (270°F) autoclaving should be extended to 4 minutes [200]. Anxieties about the low margin of safety from the 3-minute autoclaving underlie the CDC's recommendation [176] that the 3-minute cycle is not sufficient for implantable objects. The CDC did not specify any minimum duration for implantable objects, although the AAMI suggests that 10 minutes will suffice [69].
The duration of sterilization cycles at standard sterilization temperatures (121°C [250°F]) for liquids also is somewhat arbitrary. In this situation, steam is conveying only heat; thus, by extension from the standard 30-minute cycle for solid objects, longer times should be required for large volumes of liquids. The fact that most standards recommend <30 minutes or, for volumes in excess of 1 L, only 45 minutes, probably reflects the low organism burden of most such materials before sterilization.
Selection and Use of Germicides
Many physical techniques and chemical agents are useful in various contexts for disinfection or sterilization [20]. A summary of methods appropriate for various uses has been updated periodically and was last published by the CDC in 1983 [176,191]. Unfortunately, the CDC was forced to refrain from tying recommendations closely to particular products. The substituted CDC environmental guideline published in 1985 and most recently in 2003 [176] discussed disinfectants and sterilants in a more general way. Revisions of the initial CDC summary have
P.293
appeared [193,202]; they are presented in the Association for Professionals in Infection Control and Epidemiology's (APIC) “Guideline for Selection and Use of Disinfectants” written by Rutala [186]. Given the increase in type and composition of medical devices and the great variety of disinfection methods, this tabular approach is an oversimplification. The object categories overlap and are not comprehensive, and the descriptions of methods are largely references to manufacturers' recommendations for germicide concentrations, temperatures, and exposure times.
The APIC guideline [186] contains a superb discussion of the mechanisms of action, advantages and disadvantages, and tips on use of various disinfection methods. Besides the level of sterilization or disinfection, selection of a method necessitates considering its impact on the integrity of the device to be reprocessed, the possible effects on the warranty of the product and liability exposure, and occupational safety. Ethylene oxide, formaldehyde, or glutaraldehyde all pose potential risks to personnel. A proposed federal glutaraldehyde exposure standard, 0.2 parts per million (ppm) ceiling, will preclude glutaraldehyde use without evacuation hoods, personnel protection devices, or special enclosed reprocessors [203]. Whatever technique is selected, the manufacturer's instructions should be used to determine contact times and other use parameters. The disinfectant must be in contact with all relevant surfaces for the entire specified contact time.
It commonly is recommended that a particular product be purchased based on reference to standard guidelines, scientific literature, or manufacturers' recommendations. At a practical level, however, it is very difficult to use the first two. There is such a profusion of products that the standard recommendations [176,186] do not have enough specificity. Furthermore, manufacturers regularly modify their formulations and freely assert that flaws described in scientific publications [173,204] have been corrected. Thus, users are forced to rely on manufacturers' information. The only effectively regulated statement from manufacturers is that found on the label applied to the actual product.
The EPA has required that companies generate data underlying a claim that a product is a sporicide, hospital disinfectant (i.e., meets AOAC standards for disinfection of S. aureus, P. aeruginosa, and S. choleraesuis), or tuberculocide. The label has to specify the dilution, exposure time, and any other conditions required to achieve these disinfection. Reliance on the label is tricky for several reasons: the EPA or FDA can only irregularly independently verify manufacturers' claims [176,205,206], independent testing has revealed failures to meet standards [207], some disinfectant types may be inherently deficient [193], translating the EPA or FDA categories into the high-, intermediate-, and low-level disinfection system is somewhat arbitrary, and manufacturers do not always put all relevant information on the label. A manufacturer's written statement that a given product passes certain AOAC tests is important even if it is not independently verified.
Notwithstanding all of the difficulties attending reliance on the EPA-registered product label and other written statements by manufacturers, most users must depend on them in product selection. At the University of Minnesota Hospital and Clinic, sterilization is considered achievable by any germicide that passes the AOAC test as a sporicide when it is used at the dilution, temperature, and exposure time required for sporicidal activity. In addition, steam, ethylene oxide, or dry heat are believed to produce sterilization. When a liquid germicide is used for sterilization, the item to be sterilized must be rinsed in sterile water, thoroughly dried, and enclosed in a sterile package.
High-level disinfection is considered possible using any germicide that passes AOAC tests as a sporicide and a tuberculocide when used at a dilution, temperature, and time required to produce tuberculocidal activity. After high-level disinfection, rinsing in tap water followed by thorough drying and low-touch handling is permitted. In addition, high-level disinfection is considered feasible with the use of sodium hypochlorite at 10,000 ppm for 5 minutes and 1,000 ppm for 20 minutes and pasteurization at 75°C (170°F) for 30 minutes or at 90°C (195°F) for 10 minutes [208]. Intermediate-level disinfection is believed to be effected by any germicide that passes AOAC tests as a hospital disinfectant and a tuberculocide when used at the concentration and temperature required to produce tuberculocidal activity with an exposure time of ≥10 minutes. Also accepted are sodium hypochlorite at 1,000 ppm or ethanol or isopropyl alcohol at 70% to 90% at an exposure time of 10 minutes. There is no actual evidence of effectiveness for the 10-minute exposure time. Finally, low-level disinfection is considered achievable through the use of any germicide that passes AOAC tests as a hospital disinfectant when used at the label concentration and temperature required to produce hospital disinfection—sodium hypochlorite at 100 ppm or ethanol or isopropyl alcohol at 70% to 90%. There is no specified minimum exposure time for low-level disinfection; merely wiping with the disinfectant is thought to be sufficient.
Creutzfeldt-Jakob Agent
The Creutzfeldt-Jakob agent and other prions are transmissible proteins in conformations that irreversibly autocatalize conversion of the native proteins to the nonfunctional conformation. They are unusually resistant to inactivation. Transmission to humans has happened from stereotactic instruments, pituitary-derived growth hormone, corneal transplants, or dura mater grafts. Sporadic episodes in histopathology technicians have provoked anxiety about transmission to HCWs [209,210]. Critical or semicritical items previously in contact with brain tissue from Creutzfeldt-Jakob patients should be autoclaved for 1 hour at 132°C (270°F), immersed for 1 hour in 1 N sodium hydroxide, or both [211]. Although blood or spinal fluid has been found to transmit prions oxexperimentally,
P.294
there is no practical way to apply disinfecting methods to all the surfaces with which these body substances can come into contact. Given the frequency with which Creutzfeldt-Jakob disease remains undiagnosed, the potential for transmission by tissues other than brain, and the paucity of adequate disinfection methods, there is currently no practical way to accommodate fully the potential for transmission of this agent.
Environmental Sampling
Routine Microbiologic Surveillance of Inanimate Objects
Environmental sampling accounted for a large fraction of HAI control efforts in the United States through 1970. As late as 1976, 74% of hospitals with ≥50 beds conducted routine environmental culturing [212] despite explicit statements by the CDC in 1970 [213] and the American Hospital Association in 1973 [214] recommending sharp circumscription of such routine culturing. These statements advocated abandonment of the practice of routinely culturing floors, walls, linens, and air but left open the possibility of conducting epidemiologically indicated cultures, spot-checking critical hospital equipment items (e.g., respiratory care equipment), undertaking routine microbial evaluation of hospital-prepared infant formula, and verifying sterilization procedures (176). Only the last procedure, however, was deemed necessary (see Chapter 20).
Possible grounds for routine culturing of inanimate objects include HAI prevention, education of HCWs, and responding to the guidelines put forth by a number of organizations or government agencies. To contribute to HAI prevention, the culture must at least have an interpretable result. When sterility is the goal, interpretation is possible. Culturing also may be of value, however, when the need for sterility is not established (e.g., infant formula, dialysis water). Perhaps the best operational definition of interpretability is that certain results lead to specific actions. An additional, less commonly articulated criterion is that the cultured object has a high enough probability of contamination with a severe enough consequence if contaminated to justify the culture. Routine culturing of purchased sterile supplies is not justified because of the very low chance of a positive culture.
The educational value of culturing inanimate objects is limited but may be a valid adjunct to other teaching efforts. Care must be taken to prevent such efforts from growing beyond the bounds of a specific educational objective. Responding to the guidelines of various organizations quickly becomes an arcane and ineffectual exercise. First, these organizations have considerably varied standings. One must comply with the rules of regulatory agencies, such as the FDA, the federal and state Occupational Health and Safety Administration, or the federal End Stage Renal Dialysis (ESRD) program. Any hospital with a training program must meet the standards of the JCAHO. However, compliance is less essential in recommendations from the many respected government (e.g., the CDC) or non-government (e.g., the AAMI, SHEA, or APIC) agencies/organizations.
A second problem in formulating a hospital's response to these various guidelines is that the statements themselves are sometimes frankly inconsistent. For instance, the CDC's Guideline for Prevention of Intravascular Device-Related Infections [215] made no mention of culturing hospital-compounded infusion solutions when the JCAHO seemed to require such culturing. A third problem, ambiguity, is illustrated by the JCAHO statement on monitoring parenteral medication. Through 1991 (but not thereafter), solutions “manufactured” in the hospital were supposed to “be examined on a sampling basis.” We can only assume that the intent was to subject solutions to microbiologic examination.
A fourth problem is the lack of regular updating. Although some agencies, such as the JCAHO or AAMI, have instituted formal updating that includes specific rescission of previous statements, others have actually disbanded (e.g., the National Coordinating Committee on Large-Volume Parenterals (NCCLVP) sponsored by the USP and FDA [216]). A fifth complexity involves interlocking use of these dicta. The AAMI dialysis water–culturing protocol is explicitly intended to be flexible. However, the federal ESRD program requires exact compliance [217]. The American Society of Hospital Pharmacists has formally accepted the NCCLVP recommendations, giving them a longevity beyond that of their creator [216]. Despite these complexities, infection control personnel must consider the statements of these bodies in making decisions about culturing inanimate objects. If nothing else, these statements can assume substantial medicolegal importance.
A general problem that arises in considering culturing protocols for any product is determining when in the preparation-use sequence to perform the culture. It is logistically simpler to obtain the culture at the point of preparation, and the impetus to culture implements and objects used in the care of patients often comes from the quality control effort of the department preparing them. However, more relevant to patients' care is the status of the item at the time of its actual use. If cultures are positive at the end of the preparation-use sequence, efforts may be undertaken to determine the sources of contamination.
Some unusual biologic items probably should be routinely cultured. Organs, including corneas, bone, kidneys, livers, hearts, pancreases, and bone marrow for transplantation (especially if highly processed), may become contaminated in procurement, transportation, or storage. Positive cultures can have therapeutic implications in addition to suggesting the need for improvements in sterilization techniques. A biologic product that probably does not need routine culture is banked, expressed human milk intended for prompt ingestion by the donor's offspring.
P.295
It is administered orally and inevitably is frequently contaminated [218]. However, standards for donor selection, milk pasteurization, and microbiologic screening in other contexts have been published [219]. Consensus on two possible recommendations—routine culturing of hospital water for Legionella spp. and of air for fungi—may yet emerge. Routine culturing of air should be relevant only in hospitals with highly immunosuppressed patients.
Dialysis Water
The AAMI standard for culturing dialysis water (presented earlier) has been endorsed by the CDC [176] see also Chapter 23. Sampling should be conducted at least monthly. It is preferable to sample system water just before a disinfection cycle. Machine water should be taken from different machines to ensure that all defects are identified.
Hospital-Compounded Pharmacy Products
At present there is confusion regarding whether the production, mixing, or aliquoting of sterile materials by hospital pharmacies should be considered compounding or manufacturing. Preparations for individual patients clearly fall into the former category. Batches made in advance for many patients may be interpreted as being of the latter type. The implications are considerable. Compounding is governed by state boards of pharmacy and is subject to less stringent requirements. Manufacturing is regulated by the FDA and, thus, must comply with good manufacturing practice [220]. These standards, like JCAHO standards, are broadly phrased but are taken to require detailed compliance. With respect to sterility, these requirements call for culturing two items from batches of <20 items, 10% of lots of 20–200 items, and 20 units of larger lots [220]; detailed culturing procedures that include 14 days of observation for most items [220]; and quarantine of an entire batch until the sterility testing is completed [220]. These stipulations would be burdensome for hospital pharmacies.
Additional considerations apply to infusion solutions. A 1980 NCCLVP statement “endorses the concept of hospital pharmacies using sterility testing of IV admixtures as a method for monitoring the performances of pharmacy equipment and personnel” [216]. The rationale is worded to avoid the demand that sterility testing be completed before administering the solutions. The JCAHO eliminated an apparent requirement for culturing parenteral medications and solutions in 1992.
These recommendations may be challenged on several grounds. The bulk of infusion-caused infection arises from organisms ascending along the tissue-cannula interface rather than by fluid contamination (see Chapter 37). Much of the contamination of in-use infusion fluid probably arises during administration rather than compounding. Even the need for sterility of infusion fluid is arguable. Most in-use fluid contamination is present at very low concentrations, is due to relatively nonpathogenic strains, and has not been proved to be associated with illness. The least irrational program of infusion fluid culturing would focus on the in-use product, would use culture methods that do not yield positive cultures with very low levels of contamination, and would involve organism speciation to identify properly the few hazardous species capable of proliferating in the product.
Respiratory Therapy and Anesthesia Equipment
The current CDC Guideline for the Prevention of Healthcare-Associated Pneumonia [104] does not contain a recommendation for or against culturing respiratory therapy equipment. No recent organizational statement favors it. Advocates of routine culturing of breathing circuits must surmount two counterarguments: first, that there is no secure demonstration that small numbers of organisms on internal surfaces of breathing circuits cause disease and, second, that in-use breathing circuits frequently become contaminated with the patient's organisms, even if the circuits start out sterile [221]. Most reports of HAI caused by contamination of breathing circuits are not convincing. In others, it is not possible to be sure the contamination was of tubing rather than of a nebulizer or that the tubing was thoroughly dried after reprocessing [222]. Any program of routine culturing of these items should cope with the logistic problem of examining the most relevant specimens—those actually in use.
Laminar Airflow Hoods
Because HEP A filters do develop leaks, routine periodic evaluation is indicated. However, dioctyl phthalate testing is more reliable than using settling plates or other microbiologic assessments [223]. Through 1991, the JCAHO required microbiologic monitoring but in 1992 abridged the requirement to call for “a suitable area for manipulation of parenteral medications.”
Formula
Through 1977, successive editions of Standards and Recommendations for Hospital Care of Newborn Infants, published by the American Academy of Pediatrics [150], recommended routine culturing of hospital-manufactured formula obtained from nursing units. Plate counts >25 cfu/ml were deemed to indicate that the technique was faulty and immediate corrective action required. The CDC supported this measure in 1982 [176]. The 1983 and 2002 American Academy of Pediatrics and American College of Obstetricians and Gynecologists publication, Guidelines for Perinatal Care [151], has superseded the former series, and it is silent with respect to culturing hospital-manufactured formula. Similarly, subsequent statements by the American Hospital Association [149] and the JCAHO contain
P.296
no reference to this issue. Abandonment of this widely oxaccepted practice—even by those skeptical of environmental culturing [224]—probably reflects a perception that most hospitals have switched to commercially prepared formulas. Although this is true of routine infant care, there is an increase in the development of hospital-prepared specialized enteral feedings, for which specific guidelines may need to be developed.
|
TABLE 19-2 |
|||||||||||||
|
Clearly, it is necessary for infant formula and adult enteral supplements to be free of enteric pathogens and organisms capable of generating enterotoxins (e.g., S. aureus). It seems desirable that formula also be free of high concentrations of potent HAI pathogens. Neonatal oxKlebsiella oxbacteremia has followed oral ingestion of Klebsiella-contaminated breast milk [225]. Enterobacter sakazakii meningitis and death have resulted from oximbibing oxcontaminated powdered milk [226]. Freedom from Aspergillus flavus probably is desirable for all foodstuffs because of the potential of aflatoxin production. Nonetheless, previous recommendations do not call for organism identification, and it is unclear that even large numbers of organisms, excluding those mentioned previously, constitute any oxhazard.
Establishing protocols for culturing formula leads to many questions:
P.297
spores, culture techniques that promote thermophilic organism growth frequently produce excessive counts from nonhazardous formulas.
The least arbitrary routine culturing program would focus on formula to be given to the most debilitated neonates or other patients, use culturing methods yielding only human pathogens—perhaps only enteric pathogens—and be considered only a marginal supplement to general sanitary measures.
Future Issues
The recent CDC environmental guideline identified a number of critical issues for future research. Answers to many of these and other questions could enhance our efforts in improving patient care and allocating resources to ensure that we appropriately address environmental issues that place our patients at increased risk of HAIs (Table 19-2).
References
P.298
P.299
P.300
P.301