J. Scott Weese, DVM, DVSc, DACVIM
As modern medicine has advanced, our ability to save, improve and prolong life has drastically improved. Advances in medicine, surgery, critical care, oncology and other fields have had an undeniable impact on our patients, but at the same time, development of more aggressive, invasive and prolonged treatment measures has resulted in an increasing population of animals at higher risk of infection. Further, many animals are alive today that would have otherwise died in the past, but live in an immunosuppressed state and/or in close and frequent contact with veterinary hospitals. Hospitals, by their nature, are a high risk site for encountering infectious agents, and that, combined with the increased susceptibility to infection that is present in many patients, creates a high-risk environment for nosocomial infections. Nosocomial infections are an inherent risk of hospitalization, and something that can never be completely prevented. The goal, however, is to reduce the incidence of nosocomial infections, something that first requires an understanding that nosocomial infections are occurring.
Nosocomial Infections in Small Animal Hospitals
The true impact of nosocomial infections in small animal hospitals is completely unknown. Available information is largely restricted to a limited number of outbreak reports and case series from referral hospitals. Even in academic referral hospitals, data are often minimal. This is particularly true in small animal hospitals as historically much more emphasis has been placed on infection control and nosocomial infection surveillance in equine hospitals, largely in response to high profile and potentially devastating outbreaks caused by pathogens such as Salmonella.
Outbreaks receive the majority of the attention but may comprise only a small (but highly evident) fraction of nosocomial infections. The less commonly reported (and diagnosed) sporadic infections, including surgical site infections, wound infections, invasive device infections, diarrhea, and respiratory tract disease, presumably have a greater impact the health of patients that high profile outbreaks. Hospital-based transmission of concerning pathogens such as methicillin-resistant Staphylococcus aureus (MRSA), something that can occur in the absence of clinical disease, is an increasing concern in veterinary medicine from both animal and human health standpoints.
Data regarding nosocomial infections in primary care veterinary hospitals are extremely limited, something that negatively impacts development of measures to control infections (since it is difficult to control something you know little about). It is crucial to remember that the absence of data does not mean the absence of infection or risk. This author regularly consults on sporadic nosocomial infections and outbreaks in veterinary clinics, outbreaks that will never be published or published but which have a significant impact.
Despite the limited data regarding outbreaks in small animal clinics, there are reports of outbreaks caused by a variety of pathogens, including MRSA, Bordetella bronchiseptica, Salmonella, feline calicivirus and ringworm, along with non-specific syndromes such as diarrhea and respiratory tract disease. There are also numerous reports describing infection rates for various surgical procedures, something that sometimes gets overlooked as a nosocomial infection.
Primary Care Versus Referral Centres
Most reports of nosocomial infection in small animal medicine have involved referral hospitals. There are numerous possible explanations for this. The simplest is that nosocomial infections are likely more common at referral hospitals, since they may have a more susceptible patient population (more animals with advanced illness, more invasive treatments, more aggressive intensive care, more surgical procedures). Referral hospitals may also be more likely to have patients with longer hospital stays, which presumably increases the risk of an individual animal acquiring infection and developing disease during the course of hospitalization. Referral hospitals may also be larger, with more patients and staff, and therefore a greater likelihood of certain pathogens entering the facility to be spread by animals, the environment or personnel. However, another factor that must not be forgotten is that some of the nosocomial infection data almost certainly involves a reporting bias, with outbreaks more likely to be identified, investigated, characterized and published if they occur at a referral hospital.
While the relative rates of nosocomial infections between primary care and referral hospitals are unknown, it is undeniable that nosocomial infections occur in every veterinary facility. They may be rare in some facilities but they do occur, and facilities that claim to never have had a nosocomial infection are either not identifying the infections or are failing to realize that some infections that they see may have originated in the hospital.
Determining Nosocomial Versus Community-Associated Infection
Determination of the likely source of infection is important but not always easy. Some infections can occur in hospital but not become evident until after the animal is discharged. Considering the typical short hospital stay of animals in veterinary hospitals, especially primary care hospitals, a large percentage of animals that develop infections in hospital will not have clinical signs of disease until after discharge. Conversely, some animals may be admitted already infected, with clinical manifestation of infection not noted until later during hospitalization. Therefore, it is often difficult to definitively state the origin of an infection, particularly one not associated with an outbreak. However, attempts to do so are important to provide needed information about the infection control program and infection risks, as well as to determine nosocomial infection rates and to early identify outbreaks. There are no standard definitions for veterinary medicine, but crude recommendations can be made based on knowledge of different pathogens and guidelines from human medicine. A common definition is nosocomial infections are those that develop 48 hours or more after admission, although the limitations of this basic definition need to be considered. Classification of infections that develop within 48 hours is difficult and knowledge of the infectivity and incubation period of the individual pathogen is needed. Clearly, better definitions (and widespread application of them) are needed for veterinary medicine.
Barriers to Identifying Nosocomial Infections
There are numerous reasons why nosocomial infections may not be identified. Infections that develop after discharge may not be considered nosocomial, or if mild, may not be reported to the initial veterinarian. If animals are taken to a different veterinary clinic for evaluation of an infectious complication, the initial clinic may not always be notified about the infection or record that it may have been nosocomial. Even when animals develop infections in hospitals or when clear nosocomial infections are identified on recheck visits, if these events are not recorded in the medical record and if those data are not evaluated, then the occurrence of nosocomial infection may be overlooked. This can be particularly problematic in large practices. For example, if one veterinarian had five patients return with surgical site infections over a few days, it is very likely that this would be identified as a potential problem and investigated. In contrast, if five different veterinarians had a single patient return with a surgical site infection over the course of a few days, little might happen if they did not have a central way to record and communicate infection occurrences. This highlights the importance of a structure infection control program.
Consequences of Nosocomial Infections
The implications of nosocomial infection can be variable and potentially severe. Beyond the obvious potential for patient morbidity and mortality, nosocomial infections can result in significant economic impacts from restricted admissions, hospital closure, enhanced protocols and decreased caseload. Staff morale can be seriously affected from patient illness and death, negative experiences with owners of pets that develop infections, frustrations of dealing with outbreak measures and other factors. Negative publicity from outbreaks or high profile infections can result in short-term and long-term impacts on clinics. Zoonotic infections or fears of exposure to zoonotic infections by staff can also have negative consequences. Nosocomial infections with zoonotic pathogens also create additional risks if pet owners are exposed and potentially infected, as legal liability implications of such situations are unclear. Therefore, it is obvious that nosocomial infections are an important problem in veterinary medicine, and something should receive careful attention, ideally proactively to prevent, rather than try to contain, infections.
Identification of Nosocomial Infections
A critical component of any infection control program and an underused tool in veterinary medicine is surveillance. There are different forms of surveillance that can be used individually or in combination as part of a comprehensive infection control program.
Passive surveillance is a practical, efficient and cost-effective tool that can, and should, be implemented in every veterinary clinic. This form of surveillance involves the use of data that are already available, such as medical record data, admission data, bacterial culture and susceptibility testing results, and other types of diagnostic testing data. Beyond simply having these data somewhere in a medical record system, passive surveillance involves organized collection and analysis of pre-determined aspects of available data to determine elements such as endemic disease rates, antimicrobial susceptibility patterns and trends, and changes in disease patterns. Passive surveillance can also be used to ensure that unusual or particularly concerning diagnoses from a nosocomial or zoonotic standpoint (e.g., MRSA, canine parvovirus) are rapidly identified and conveyed to relevant individuals so that appropriate measures can be taken.
As with any aspect of surveillance, simply collecting data is of no use if no one analyses it or if there is no plan for using the results. Therefore, passive surveillance must be an organized, structured activity with clear objectives and methods.
In the absence of an ongoing infectious disease outbreak, passive surveillance is likely adequate for most clinics. An example of passive surveillance would be monitoring surgical site infection rates, either involving overall procedures or, ideally, overall rates plus specific rates for common or high-risk procedures. Critical limits for specific hazards can be established, with a pre-determined plan of action should limits be exceeded.
Another passive surveillance activity that can be useful involves collection and analysis of routine culture and susceptibility testing results. Having data about bacterial species isolated from various sites and their susceptibility patterns can be useful in selection of empirical therapy when culture is not possible or while culture results are pending. The more detailed characterization of surveillance, the better, as this allows for more specific and appropriate application of data. For example, pathogen and resistance patterns may be different in SSIs originating in hospital versus the community, and knowing culture result patterns from both origins would facilitate a more direct empirical approach.
While passive surveillance can be relatively easy and cost effective, people using the data must understand any limitations because they it is highly dependent on the quality of available data and the potential for biases in the data. For example, if cultures are usually only submitted for recurrent infections or those that have failed initial therapy, there will probably be a bias towards more resistant pathogens. This does not negate the value of the data; it just means that potential limitations must be considered.
Active surveillance involves gathering data specifically for infection control purposes. Therefore, it is usually more expensive and time consuming but if properly done, can provide the highest quality data. Active surveillance can involve patients, medical equipment or the environment. Active surveillance of patients usually involves screening of patients for infection or colonization with specific pathogens. As with other aspects of surveillance, active surveillance must be performed using an organized, pre-determined program designed to fulfill specific objectives, with associated plans for analysis of data and use of results. Broad, routine active surveillance is rarely indicated in veterinary hospitals and is typically reserved for large facilities with increased infection control threats and personnel available to direct such testing, or during a specific outbreak investigation. Even then, it is of limited routine use. An example of active surveillance is collection of nasal and rectal swabs from all animals being admitted to a hospital, whether or not they have signs of infection, to screen for methicillin-resistant Staphylococcus aureus as is commonplace in human medicine and as has been described for equine patients. This typically involves screening of patients (all patients or selected high-risk groups) at the time of hospital admission to identify carriers. Screening allows for isolation of carriers and decrease risk of transmission to other patients or healthcare personnel. It also may indicate patients at increased risk for developing clinical MRSA infection. Therefore, this type of surveillance clearly has specific objectives, a defined plan and a pre-determined response to the results. The author is unaware of any institutions that perform ongoing MRSA (or other pathogen) surveillance in small animals for infection control (not research) purposes, and of no evidence indicating that routine active surveillance is currently indicated.
Syndromic surveillance involves the analysis of certain clinical signs that may be indicators of an infectious disease. It is often used as a measure to identify outbreaks in complex populations that are not easy to monitor otherwise. For example, monitoring work or school absences or over-the-counter drug prescriptions can be used as an indication that a disease outbreak might be present in people in a community. This type of surveillance does not indicate what disease outbreak is present or even confirm that an infectious disease is the cause of the syndrome, but it provides readily accessible data to act as an easy or early warning and lead to specific investigation. Because this type of surveillance involves syndromes, not specific diagnosis, it is easier to implement and does not require supporting laboratory testing or detailed clinical interpretation. In contrast, because of its broad and non-specific nature, patients with the same syndrome can have unrelated disease caused by different pathogens or have a non-infectious condition. Therefore, syndromic surveillance has a lack of specificity but, as long as the limitations are understood, syndromic surveillance can be a useful surveillance tool.
In veterinary hospitals, syndromic surveillance can be used as a means to rapidly and easily identify potential indicators of the presence of an infectious disease, particularly nosocomial disease. Development of signs such as fever, diarrhea, coughing or vomiting, while certainly not always related to an infectious disease, can be early and readily identifiable indicators of infectious diseases. Because these signs are obvious and can be easily detected and recorded, syndromic surveillance can be a practical tool to detect infections or early stages of an outbreak. Syndromic surveillance is only a means to identify problems for further investigation, and any potential problems must be investigated further through other means.
A similar approach can be used as a preventive measure through identification of animals with potentially infectious diseases prior to entering the hospital. If front-office staff or people taking initial appointment calls can identify animals that may be infectious, these animals can be handled differently. Identification of animals with complaints such diarrhea, coughing/sneezing, acute onset of neurological disease, or wound infection before arrival can allow for consideration of how they should be managed upon arrival (i.e., owner interviewed without the animal initially, owner and animal met outside, animal admitted directly to an examination room or isolation), and therefore potentially reduce the risk of exposure of other patients, owners or staff, or contamination of the hospital environment. By its broad nature, syndromic surveillance is not highly specific and non-infectious animals will be identified and managed with unnecessary precautions. However, considering the minor implications of over-identifying potentially infectious cases versus the potential problems with failing to identify an infectious case, lack of specificity is acceptable.
Environmental surveillance is a commonly used, and often improperly used or minimally useful surveillance measure. There is an inverse relationship between the availability of specialized infection control personnel in human hospitals and the amount of environmental surveillance, with environmental surveillance decreasing in facilities with trained infection control personnel and both the CDC and American Hospital Association advocate against routine environmental surveillance. Yet, environmental culture is often the first response by veterinary clinics in response to a perceived nosocomial infection problem, likely a reflection of the lack of available information and poor understanding of general infection control practices. Efforts such as sampling various environmental surfaces, inanimate objects and air in the absence of a specific objective and organized approach with a pre-determined action plan are simply an inadequate use of time and resources. The hospital environment is not, and never will be, a sterile environment. Contamination with various opportunistic pathogens is common and completely expected. Routine cleaning and disinfection practices are designed to reduce the environmental burden, with greater emphasis on higher risk items or areas but sterilization of the environment in not a goal. Therefore, identification of bacteria in general hospital environmental sites is not particularly useful.
Infection control is an important and often overlooked component of good veterinary practice. Undoubtedly, nosocomial infections occur in every veterinary practice, regardless of type or size. The relative incidence and risk of nosocomial infections may vary between practices but all practices need to be aware of the risks and implement basic infection control measures to detect, prevent and control nosocomial infections. As the awareness of nosocomial and zoonotic infections increases, and new infectious disease challenges emerge, it is clear that a more formal approach to infection control is required.
References are available upon request.