Given the paucity of supporting data, it is unclear how similar the subgroups lumped together under the HCAP umbrella really are. Intuitively all of these patients are at increased risk for colonization with resistant pathogens, given their frequent contact with the healthcare system and/or its extenders. The repeated observation that the presence of any one of the factors defining HCAP—recent admission, nursing-home-residence, renal failure, and so on—is associated with increased mortality in patients with nosocomial infections strengthens this assumption.17-19
There are data that challenge the conclusion that such nosocomial pathogens are a frequent enough cause of pneumonia in these patients to merit routine empiric therapy, however. In one large retrospective database review, hemodialysis patients admitted for pneumonia had a spectrum of causative pathogens that was a hybrid of community- and nosocomial-acquired microbes.20 Similarly, in a large systematic review of nursing-home acquired pneumonia studies published between 1978-1994, the causative pathogens were a diverse blend of community- and nosocomial-acquired organisms.21 Both of these studies were retrospective analyses that were limited by the availability and quality of the previously collected microbiology data.
The only study looking specifically at the epidemiology of HCAP as defined by the ATS-IDSA guidelines was published in December 2005—10 months after the guidelines themselves.22 This study raises the possibility that earlier epidemiology studies of pneumonia may suffer from obsolescence, given recent trends in worsening microbial resistance and an increasingly complex medical system.
The authors of this large (n=4,453), retrospective, multi-center database analysis concluded that HCAP was justified as a new category of pneumonia based on the observation that the pathogens causing HCAP were more similar to HAP and VAP than they were to CAP. For instance, in HCAP, Staphylococcus and Pseudomonas species were isolated in 46.7% and 25.3% of patients, while in CAP patients, these organisms were recovered in 25.5% and 17.1%, respectively. Conversely, Streptococcus pneumoniae and Haemophilus species predominated in CAP (16.6% and 16.6%) but were less common in HCAP (5.5% and 5.8%), HAP (3.1% and 5.6%), and VAP (5.8% and 12.2%). Compared with CAP HCAP was also associated with more severe disease, a higher mortality rate, a prolonged length of stay, and significantly increased costs.
Although not particularly germane to hospitalists, these recommendations were also found to be problematic for patients who were not hospitalized. Even for individuals residing in chronic-care facilities, the use of multiple broad-spectrum agents being given by IV would pose a logistic disaster that would quickly overwhelm staff. Safety issues would also likely abound, given the need for IV lines, pumps, and tubing in a patient population that would have a high density of cognitive deficits and a relatively unfavorable nurse-to-resident ratio. Further, the concept of empiric broad-spectrum therapy mandates subsequent de-escalation based on microbiology specimens obtained at the time of initiating antibiotics. It has been repeatedly illustrated that obtaining high-quality specimens outside the hospital setting is difficult. As a result, most non-hospitalized patients would be started on broad-spectrum therapy without any realistic possibility for de-escalation, a situation that would only accelerate the selection pressure on resistant organisms.
The guidelines are also vague regarding whether Legionella should be empirically treated in patients with HCAP. While this pathogen has long been accepted as a common and under-diagnosed cause of community-acquired pneumonia, the epidemiology of Legionella pneumophila (Legionnaires’ disease) in nosocomial pneumonia is more controversial. Although it is likely a relatively common cause of endemic nosocomial infections, the initial reports of Legionnaires’ disease have left the misperception that this pathogen is only a cause of epidemic disease.23-26 In the 2005 ATS-IDSA guidelines, Legionella is listed as a “potential pathogen.” Yet it would only be treated if a macrolide (e.g., azithromycin) or a fluoroquinolone (e.g., ciprofloxacin or levofloxacin) were prescribed.