SPECIAL COMMENTARY

Healthcare-associated pneumonia is a heterogeneous disease,

and all patients do not need the same broad-spectrum antibiotic
therapy as complex nosocomial pneumonia
Veronica Britoa and Michael S. Niedermana,b,c
a
Department of Pulmonary, Critical Care and Sleep Purpose of review
Medicine, Winthrop-University Hospital, bDepartment
of Medicine, SUNY at Stony Brook and cDepartment
Healthcare-associated pneumonia (HCAP) develops in patients who have recently had
of Medicine, Winthrop-University Hospital, Mineola, contact with nosocomial and drug-resistant pathogens, because of a history of
New York, USA
hospitalization in the past 90 days, need for hemodialysis or home wound care, or
Correspondence to Dr Michael S. Niederman, residence in a nursing home. HCAP was included in the 2005 American Thoracic
Chairman, Department of Medicine, Winthrop-
University Hospital, 222 Station Plaza North, Suite Society/Infectious Diseases Society of America guidelines for nosocomial pneumonia,
509, Mineola, NY 11501, USA with the recommendation that all such patients receive empiric therapy with a multidrug
Tel: +1 516 663 2381; fax: +1 516 663 8796;
e-mail: mniederman@[Link] regimen directed against drug-resistant organisms. The purpose of this review was to
examine articles published since the guidelines were developed to see whether this
Current Opinion in Infectious Diseases 2009,
22:316–325
therapy recommendation is correct.
Methods
All articles published since July 2004 were identified using PubMed and the key words
HCAP, nursing home-acquired pneumonia, and antibiotic therapy. The search was
limited to adults, with a focus on clinical trials, reviews, meta-analyses, or practice
guidelines.
Recent findings
We identified eight unique studies of HCAP, which were either prospective or
retrospective series, with bacteriologic data on both Gram-negative and Gram-positive
organisms. We also examined three prospective, randomized therapy trials of nursing
home-acquired pneumonia that included limited bacteriologic data. We found that
patients with HCAP were a heterogeneous group, with some at risk for multidrug-
resistant organisms, and others not, and this accounted for the observation that many
patients were successfully treated with monotherapy regimens or with regimens used
for patients with community-acquired pneumonia. Patients at risk for multidrug-resistant
pathogens were those with severe illness or those with other risk factors including:
hospitalization in the past 90 days, antibiotic therapy in the past 6 months, poor
functional status as defined by activities of daily living score, and immune suppression.
Conclusion
On the basis of the risk factors identified in recent studies, we developed an algorithm
for empiric therapy of HCAP, which suggests that not all such patients require
a broad-spectrum multidrug regimen in order to achieve appropriate and effective
therapy. This algorithm needs validation in future studies.

Keywords
antibiotic resistance, bacteriology, healthcare associated, nursing home, pneumonia,
therapy

Curr Opin Infect Dis 22:316–325

ß 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins
0951-7375

(HAP) or nosocomial pneumonia [1]. HCAP included

Background patients who developedpneumonia in or outof the hospital,
In 2005, in a joint statement by the American Thoracic but who had a history of contact or exposure to the health-
Society (ATS) and the Infectious Diseases Society of care environment, which meant that they were potentially
America (IDSA), healthcare-associated pneumonia infected with multidrug-resistant (MDR) bacteria. The
(HCAP) was defined and included within therapy recom- risk factors for HCAP were identified as: hospitalization
mendations as a form of hospital-acquired pneumonia for 2 days or more in the preceding 90 days, residence in a
0951-7375 ß 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins DOI:10.1097/QCO.0b013e328329fa4e

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Healthcare-associated pneumonia: special commentary Brito and Niederman 317

nursing home or extended care facility (nursing home- ology of patient populations, or both, included within the
acquired pneumonia, or NHAP), receipt of home infusion definition of HCAP.
therapy (including antibiotics), chronic dialysis within
30 days, home wound care, and exposure to a family
member infected with a MDR pathogen. Other studies Inclusion/exclusion criteria
[2] of HCAP have included patients with immune compro- As a background, the recovered studies were reviewed to
mise, but this group was excluded from the ATS/IDSA provide perspective on the problem of HCAP bacteri-
definition. One potentially at-risk population that was not ology and therapy. However, only studies published since
mentioned is the healthcare worker, especially those July 2004 were included in the detailed analysis in order
with chronic medical illnesses who care for patients to update the database that was used to develop the 2005
infected with MDR pathogens. ATS/IDSA guidelines. Included studies were retrospec-
tive or prospective observational cohorts of patients
The ATS/IDSA guideline stated that all patients with included within the category of HCAP. Prospective stu-
HCAP should receive empiric therapy directed at MDR dies of therapy were also included, but none enrolled a
pathogens, including methicillin-resistant Staphylococcus mixed HCAP population, and all only involved nursing
aureus (MRSA) and Pseudomonas aeruginosa. This means home patients. Although some studies included immune-
that all HCAP patients would require parenteral therapy suppressed populations, any study dealing exclusively
with usually three different antibiotics. The recommen- with immune-suppressed patients was excluded.
dation was based on a limited number of studies available Abstracts, not published in full, were not considered.
at the time, primarily from the USA, but since then, other
studies from other countries, as well as the USA, have
reported a different picture of HCAP, with a lower Data extraction and synthesis
frequency of MDR pathogens calling into question the Studies were examined to define the bacteriology of
need for routine broad-spectrum empiric therapy [2–8]. HCAP, site of care, therapy used, and the relationship
In addition, in North America, it is clear that not all HCAP of outcome to therapy. The search strategy yielded six
patients are severely ill or even hospitalized when they studies [2–5,7,10] of HCAP that included more than just
develop pneumonia, and these patients, generally residing nursing home patients, three were retrospective and
in nursing homes, can receive oral antibiotic therapy with observational with no randomized intervention, and
single agents with great success, somewhat contrary to the two of these three involved the same patients. Two
recommendations of the guideline [9]. studies [5,7] were prospective observational reports,
one comparing HCAP with community-acquired pneu-
The purpose of this review is to examine the current data monia (CAP) and one comparing NHAP with nosocomial
published after the literature review used for the guide- pneumonia. Another study used case–control method-
lines to better understand the population of patients with ology, but whereas the latter included HCAP with
HCAP, their associated bacteriology, therapy, and the other healthcare-associated infections, it only examined
success of therapy, in order to recommend an updated patients infected with MRSA [10]. Studies were from
approach for antibiotic management. North America (three studies), Europe (two studies), and
Japan (one study). Two of the North American studies
[2,4] involved the same population, but each study eval-
Search strategy uated different aspects of bacteriology and therapy. We
Studies of HCAP therapy were evaluated by searching also evaluated four other studies [6,8,11,12] in which the
PubMed in September 2008. Initially, when the term only HCAP patients evaluated were those with NHAP,
HCAP (614 studies) was combined with the term anti- but all of these were prospective and observational, two
biotic therapy (234 208 studies), using the word ‘and’, a from Europe and two from the USA. The studies from
total of 160 studies remained. When the term nursing the USA [11,12] were the only ones to systematically
home pneumonia was combined with the term antibiotic use quantitative bronchoscopic culture sampling, and
therapy, there were 168 studies. These searches were there was some overlap in the included population of
limited to adults, clinical trials, reviews, meta-analyses, or patients.
practice guidelines. To broaden the search, the term
hemodialysis was combined with the term pneumonia We combined the eight unique HCAP study populations
to yield 403 studies, and when the term hemodialysis was that were evaluated either prospectively or retrospec-
combined with the term pneumonia bacteriology, there tively and that had bacteriologic data on both Gram-
were 70 studies. positive and Gram-negative bacteria into Tables 1
and 2 [2,3,5–8,11,12]. Study size varied, with the
This database of studies was reviewed and cross-refer- North American studies having 75 and 135 and 431
enced to identify original studies of therapy or bacteri- and 988 HCAP patients, whereas the European studies

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Table 1 Unique studies of healthcare-associated pneumonia published since July 2004
318 Evidence-based medicine

Number of patients/
percentage with Inappropriate Atypical
Study Design positive cultures therapy S. pneumoniae S. aureus Gram negatives H. influenzae pathogens

El Solh et al. [11] Prospective, single-center 98/68% 25% 31% 28% 2% 1 and 2% with
(NHAP) study in the USA influenza A
Martinez-Maragon Prospective, single-center 25/24% 19% (treatment 33% 50% 17%
et al. [6] (NHAP) in Spain modifications of
initial empirical
antibiotic)
Kollef et al. [3] Retrospective, multicenter 988/100% 5.5% 46.7% 59.5% 5.8% (Haemophilus
in the USA (all patients spp.)
with positive cultures)
Carratala et al. [5] Prospective, multicenter 126/67.5% 5.6% 27.8% 2.4% 4% 11.9% 1.6%
study in Spain
Micek et al. [2] Retrospective, single center 431/100% 28.3% 10.4% 44.5% 25.5% 4.2% (Haemophilus 0.2% (Legionella
in the USA (all patients (P. aeruginosa) spp.) spp.)
with positive cultures)
El Solh et al. Prospective, single center in 75/65% 14.3% 26.5% >47% 48.1% Not reported
[12] (NHAP) the USA – only severe
(mechanically ventilated)
pneumonia
Kothe et al. [8] Prospective, multicenter in 205 Not reported 2.3% 18.8%
(NHAP) Germany
Muruyama Prospective, single center 75/72% 33.3% 4% 4% 34.7%
et al. [7] in rural Japan (C. pneumoniae)
C. pneumoniae, Chlamydophila pneumoniae; H. influenzae, Haemophilus influenzae; NHAP, nursing home-acquired pneumonia; P. aeruginosa, Pseudomonas aeruginosa; S. aureus, Staphylococcus
aureus; S. pneumoniae, Staphylococcus pneumoniae.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Healthcare-associated pneumonia: special commentary Brito and Niederman 319

Table 2 Heterogeneity of the populations included in unique studies of healthcare-associated pneumonia published since July 2004
Percentage with
either S. aureus or enteric
Gram negatives (some
Study Design patients had both) Limitations in the population studied

El Solh et al. [11] Prospective, single center in the 59% Only nursing home patients, all admitted to the
(NHAP) USA ICU; excluded anyone admitted to the hospital
in the preceding 6 months; excluded immune
suppressed
Martinez- Maragon Prospective, single center 67% Only nursing home pneumonia. None
et al. [6] (NHAP) in Spain admitted to ICU
Kollef et al. [3] Retrospective, multicenter in the Almost all Only patients with positive cultures; information
USA (all patients with from a large database. 49.6% from a nursing
positive cultures) home. 24.1% mechanically ventilated
Carratala et al. [5] Prospective, multicenter evaluation 6.4% 43.7% with hospitalization in the preceding
in Spain 90 days. 25.4% from nursing home. 6.3%
admitted to ICU, 3.2% mechanically ventilated
Micek et al. [2] Retrospective, single center in 70% Only enrolled those with a positive respiratory
the USA (all patients culture. Included immune suppressed, anyone
with positive cultures) hospitalized in the past 12 months (93.3%
had recent hospitalization, including 69% in
the past 90 days)
Only 28.1% from nursing home
48.7% in ICU, 44.5% mechanically ventilated
El Solh et al. [12] Prospective, single center in the >75% All were mechanically ventilated. Excluded immune
(NHAP) USA – only severe (mechanically suppressed, antibiotic therapy in the preceding
ventilated) pneumonia 72 h, and those with witnessed aspiration
Kothe et al. [8] Prospective, multi center in 21.1% Only nursing home pneumonia evaluated, but
(NHAP) Germany excluded any patient with recent hospitalization
in the preceding 28 days
Muruyama et al. [7] Prospective, single center in 8% Only nursing home pneumonia. Exclude immune
rural Japan suppressed and hemodialysis patients. 8.5%
admitted to ICU
NHAP, nursing home-acquired pneumonia; S. aureus, Staphylococcus aureus.

had population sizes of 25, 126, and 205 HCAP patients, bacteriology. In general, the studies that included mainly
and the Japanese study had 75 HCAP patients. An patients who had recently been hospitalized or those with
additional retrospective study [13] was included, but it severe NHAP had a higher frequency of MDR pathogens
only focused on patients with pneumonia and hemodia- than the studies that included HCAP patients who were
lysis and provided limited bacteriologic data with no less severely ill or who met this definition for reasons
comment on therapy. other than recent hospitalization. These findings are
consistent with an algorithm presented in one of the
Three additional prospective therapy trials were also NHAP studies by El Solh et al. [11]. In that study [11],
evaluated, but only involved HCAP patients who came there were 88 patients with severe NHAP, but only 17
from nursing homes (NHAP). In one study [9], 680 had drug-resistant pathogens, and they were individuals
patients were evaluated, some treated in the hospital who, in addition to severe illness, had either a history of
with intravenous or oral therapy, whereas others received antibiotic therapy for at least 3 days in the past 6 months,
oral therapy in the nursing home. In one of the other a poor functional status (defined by activities of daily
studies, 69 patients were treated parenterally in the living), or both.
nursing home, whereas in the other study, 51 were treated
intravenously in the hospital [14,15]. Another study [16] Micek et al. [2] retrospectively studied the bacteriology
was reviewed but not analyzed in detail, because even of HCAP and CAP in 639 patients at a single US hospital
though it examined 170 nursing home pneumonia over a 3-year period (2003–2005). Unlike some other
patients admitted to the hospital, it provided very limited HCAP studies, all pneumonia patients had positive
data about bacteriology and the specific antibiotic bacteriology, with cultures collected from sputum, tra-
therapy used. cheal aspirate, bronchoscopy samples, or blood cultures.
HCAP was defined similar to the ATS/IDSA guidelines
with the exception that the investigators also included
Results those hospitalized in the past 12 months (rather than
The studies included in Tables 1 and 2 incorporated a 3 months) and also those with immune suppression
heterogeneous group of patients, with a wide range of (steroid therapy of at least 5 mg prednisone/day, HIV

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 320 Evidence-based medicine

infection, transplant, or other immune suppression such retrospective analysis, whereas the study by Carratala
as with chemotherapy). Using this definition, HCAP was et al. [5] was prospective; the study by Micek et al. [2] only
more common than CAP (431 versus 208 patients), and included patients with positive cultures, thus limiting
the majority of patients were defined as having HCAP the type of patients enrolled, and the studies were done
because of recent hospitalization (in 93.3%), with 69% in different institutions. This latter fact explains the
having been hospitalized in the past 90 days [2]. In the heterogeneity of the populations with HCAP that have
HCAP population, 48.7% were admitted to the ICU and been included in studies. Although most of the patients
44.5% were mechanically ventilated. The most common in the study by Micek et al. [2] were recently hospital-
pathogens for the entire population (CAP and HCAP) ized, the population in the Spanish study was different.
were MRSA (24.6%), S. pneumoniae (20.3%), P. aeruginosa In that study, only 55 of the 126 HCAP patients (43.7%)
(18.8%), and Methicillin-susceptible S. aureus (13.8%). had been admitted to an acute care hospital for 2 or more
However, MRSA was more common in HCAP than in days in the 90 days, whereas 40 (31.7%) had HCAP
CAP (30.6 versus 12.0%), as was P. aeruginosa (25.5 versus because of attending a hospital or a hemodialysis clinic
4.8%), whereas S. pneumoniae and Haemophilus species or receiving intravenous chemotherapy in the preceding
were more common in CAP. Patients with HCAP had a 30 days; 32 (25.4%) resided in a nursing home or a long-
significantly higher mortality rate than CAP (24.6 versus term care facility, and 18 (14.3%) received home health-
9.1%, P < 0.001), and the administration of inappropriate care. Nineteen patients had more than one reason to be
initial therapy (use in the first 24 h of an agent that was included in the HCAP group. In addition, in the Spanish
not active against the identified pathogen) was more study, only 6.3% of the HCAP patients were admitted to
common in HCAP patients than in CAP patients the ICU and 3.2% were mechanically ventilated com-
(28.3 versus 13.0%, P < 0.001). No specific information pared with 48.7 and 44.5%, respectively, in the study by
was given about antibiotic therapy, except the finding Micek et al. [2].
that when patients received a CAP therapy regimen in
the setting of HCAP, therapy was inappropriate 30.9% The Japanese study [7] prospectively compared 75
of the time. The organisms in HCAP patients that patients with NHAP with 33 patients with HAP. The
were treated with inappropriate therapy most often were study was done in a rural area that has over 30% of the
P. aeruginosa, MRSA, and other nonfermenting Gram population above the age of 65 years. All the patients with
negatives. In a multivariate analysis, the relevant risk NHAP had been institutionalized for at least 5 days prior
factors for mortality were: mechanical ventilation [odds to the onset of pneumonia. Immune-suppressed and
ratio (OR), 5.05], bacteremia (OR, 3.26), and inappropri- dialysis patients were excluded from this study. Only
ate therapy (OR, 2.19). 8.5% of the NHAP patients were admitted to the ICU.
The mortality rate of NHAP was high at 37.3%, but still
The findings by Micek et al. [2] differ from the data of lower than for HAP. In spite of this high mortality, there
Carratala et al. [5], a Spanish study which compared 126 was a relatively low frequency of MDR pathogens, with
HCAP patients with 601 CAP patients. Although they only 4% having S. aureus and 5.3% having enteric Gram
showed a higher frequency of inappropriate therapy and negatives. In the study, 72% of the NHAP patients had an
mortality in HCAP than in CAP patients, the mortality etiologic diagnosis, and the surprising finding in this
was much lower than in the study by Micek et al. [2], and study was the high frequency of atypical pathogens,
there was also a much lower frequency of MDR patho- which were found in 37.3% of the NHAP population,
gens in HCAP patients than was found in the study by primarily Chlamydophila pneumoniae, diagnosed serologi-
Micek et al. [2]. In the study, both MRSA and enteric cally. The high frequency of atypical pathogens contrasts
Gram negatives were more common in HCAP than in with the other studies of HCAP, but this was the only
CAP patients, but in the HCAP patients, the frequency study to systematically look for evidence of their pre-
of S. aureus was only 2.4%, and the frequency of enteric sence. Only about half of the NHAP patients received
Gram negatives was only 4%. The therapy used for a macrolide or minocycline, and although those with
these patients reflected the bacteriology. Overall, 75% Chlamydophila who were appropriately treated had a
received monotherapy, 62% with a b-lactam and 10% 71% response rate, mortality was not affected by this
with a quinolone. When a combination regimen was appropriate initial therapy, raising a question about the
used, it consisted of a b-lactam and quinolone in 21% importance of this organism. The therapy of NHAP was
and b-lactam and a macrolide in 3%. Thus, many patients more often dual therapy than in the HAP population, but
did receive a CAP regimen, but the frequency of inap- the HAP patients tended to receive broader spectrum
propriate therapy was only 5.6%. drugs with more carbapenem and less cephalosporin use
than the HCAP patients.
The disparate findings between these two studies may be
related to differences in study design and in the popu- The other HCAP studies were not as well detailed or
lation enrolled (Table 2). Micek et al. [2] performed a did not include as heterogeneous a population. Kollef

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Healthcare-associated pneumonia: special commentary Brito and Niederman 321

et al. [3] retrospectively analyzed a large database of validated a prediction model for the presence of MDR
4543 pneumonia patients of whom 988 had HCAP. In pathogens. All patients had severe pneumonia, but
this population, 49.6% of patients with HCAP came MDR pathogens were absent from those with no
from a nursing home, and 24.1% were mechanically history of prior antibiotics and with good functional
ventilated, but there is no mention of the percentage status (based on activities of daily living score). The
who were admitted to the ICU [3]. The mortality of frequency of MDR pathogens was highest in those with
HCAP was similar to that of nosocomial pneumonia, both risk factors, and those who received prior anti-
higher than that of CAP, and lower than ventilator- biotics as the only risk factor had a higher frequency of
associated pneumonia. The most common organism in MDR pathogens than those with poor functional status
the HCAP population was S. aureus, present in 46.7% as the only risk factor. For the 135 patients, the
(57% of these were MRSA), and P. aeruginosa (25.3%). sensitivity of the model for predicting MDR pathogens
This study is limited by its retrospective dependence was 100%, with a specificity of 58.9%. No therapy
on a database with bacteriologic data of uncertain information was given in this study.
accuracy. The French study [10] was a prospective
case–control evaluation of the risk factors for health- The second study by El Solh et al. [12] did report
care-associated MRSA infections that were diagnosed extensive microbiology on 75 patients with NHAP
within 48 h of admission. Of the 175 MRSA cases, only (excluding those with immune suppression), all of whom
41 were pneumonia, and patients and the HCAP risk were mechanically ventilated, but no details of therapy or
factors were defined for the entire group, not just the outcome were provided. However, this study is the only
pneumonia patients. These risks were: hospitalization one to systematically use quantitative bronchoscopic
within the past 3 years (85%), home nursing care within cultures, both protected specimen brush (PSB) and
the past 3 years (65%), transfer from another healthcare BAL. They found organisms present above a quantitative
facility (25%), family contact with the healthcare threshold in 49 of the patients, with 56 different micro-
environment (52%), and antibiotic use in the past organisms. S. aureus was most common (n ¼ 13), followed
6 months (53%). This study is obviously limited by its by S. pneumoniae (n ¼ 7) and Escherichia coli (n ¼ 7).
study of only MRSA, but it did identify risk factors for Although other Gram negatives were identified (Serratia
healthcare-associated MRSA infection that were: prior marcescens and Proteus mirabilis), only three bronchoscopic
receipt of home nursing care (which showed an samples showed P. aeruginosa, and Acinetobacter was not
exposure–frequency relationship), prior hospitalization, reported. Interestingly, the authors found that in this
transfer from another institution, and age of more than intubated population, quantitative endotracheal aspirates
65 years. correlated well with the results from quantitative
bronchoscopy.
The four studies of NHAP give additional data about
bacteriology, but are limited by the populations studied The other study [8] to include NHAP was a prospective
and the data provided. The Spanish study [6] evaluated study of 2647 CAP patients that compared older patients
only 25 hospitalized patients with NHAP, and only six with younger individuals, but did include within the
had an etiologic diagnosis, which was mostly based on study 205 individuals coming from nursing homes who
blood cultures. Although none were admitted to the would now be classified as having HCAP. Compared with
ICU, 28% died, and the identified organisms were: S. CAP patients, those with NHAP had a higher incidence
pneumoniae in two, S. aureus in three, and enteric Gram of Gram negatives (18.8%), but the incidence of S. aureus
negatives in one. The specific therapies used were not (2.3%) and P. aeruginosa (1%) was relatively low. The
specified, but initial therapy was changed in 16% of the presence of NHAP increased mortality by four-fold com-
patients with NHAP. El Solh et al. [11,12] did two pared with elderly CAP patients living in the community.
studies of NHAP patients, all with severe illness Specific therapy was not described, but 39.7% had some
requiring mechanical ventilation and admitted to the change in antibiotics, with 14% having a change due to
ICU. In the first study [11], 135 patients were evalu- ineffectiveness and 2% due to resistance. In contrast to
ated, 88 to derive a prediction rule for bacteriology and some of the other studies, the investigators excluded
47 to validate the rule. The study excluded patients any NHAP patient with a recent hospitalization (within
who had been hospitalized in the past 6 months and 28 days).
those with immune suppression. Of the 135 patients, 93
had a microbial cause defined by extensive diagnostic Hemodialysis patients are also included within the
testing, including protected bronchoalveolar lavage HCAP definition, but there are few studies of pneumonia
(PBAL). The organisms found were: S. aureus in in this population. One study [13] retrospectively ana-
31% (20% of all patients had MRSA), enteric Gram lyzed a Medicare database of 10 635 dialysis patients and
negatives in 28% (9% of all patients had P. aeruginosa), provided limited data about pneumonia bacteriology. In
and S. pneumoniae in 25%. The authors derived and the study, of the 3101 pneumonia hospitalizations, no

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 322 Evidence-based medicine

cause was identified in 82%. When an organism was

Discussion
found, Gram negatives outnumbered Gram positives
Many issues enter into the decision about initial empiric
almost two to one (11.1 versus 4.8% of all). No comments
therapy of HCAP, and the patients with HCAP are quite
were made about therapy. In this group, staphylococcal
heterogeneous with differences in severity of illness
species were present in 0.4% and P. aeruginosa in 2.8% of
(mechanically ventilated or not), site of care (in or out
the entire population (recognizing that no pathogen was
of the hospital), route of therapy (oral, intravenous, or
reported in over 80% of the patients).
intramuscular), and the underlying reason for being
defined as having HCAP. These differences in study
There were three prospective, randomized trials of
design are summarized in Table 2. In addition, some
therapy for NHAP that give some insight into the types
HCAP patients, particularly those in nursing homes, are
of therapy that are effective for different subpopulations
at risk for infection with community-based pathogens
of patients. One study [9] from Canada included 680
such as Legionella, C. pneumoniae, and influenza virus.
patients who were randomized to usual care or a clinical
Because of these factors, some patients are at risk for
pathway that attempted to administer oral quinolone
pneumonia with MDR pathogens, whereas others are
therapy and keep the patient from going to the hospital,
not, and this can explain why some have been effectively
if possible. Only 110 of the 680 patients were ever
treated with antibiotic regimens that do not target these
hospitalized, making this a less severely ill population
organisms, whereas others have received a multidrug
than the other studies. There were no bacteriologic data
regimen. For all of these reasons, not all patients with
from this study, but the majority of patients were success-
HCAP routinely need the same broad-spectrum anti-
fully treated with levofloxacin, usually orally, with good
biotic regimen as patients with complex nosocomial
outcomes implying the lack of MDR pathogens in most
pneumonia, as suggested in the ATS/IDSA HAP guide-
patients. The other therapy studies were prospective
lines [1]. When the ATS/IDSA guidelines suggested that
randomized, comparative evaluations of two different
HCAP should be treated similar to HAP, with a focus on
therapies for NHAP.
MDR pathogens, the implied, but not stated, assumption
was that the patients being evaluated were those in the
One was a prospective, double-blind, randomized
hospital who would be treated with intravenous anti-
study of 74 hospitalized patients with HCAP, with
biotics, and this assumption did not account for the
33 receiving intravenous monotherapy with ertapenem
heterogeneity of HCAP populations. We believe that
compared with 41 receiving intravenous therapy with
the proper approach to empiric therapy of HCAP is
cefepime, but patients with pseudomonal risk or
complex, necessitating the definition of subpopulations,
severe illness (none were intubated) were excluded
some of which should be managed similar to nosocomial
[15]. Both monotherapies were highly effective for
pneumonia, some similar to CAP, and some with a hybrid
patients with NHAP, being 90% with cefepime and
approach, both in the choice of therapeutic agents as well
75% with ertapenem. These results occurred even
as in the site of care and route of antibiotic administration.
though there were many patients with potentially
MDR pathogens, although the study also included
patients with traditional nosocomial pneumonia, and The data reviewed not only highlight the complexity of
the bacteriologic data were not separated from those HCAP, but also demonstrate that for many patients, the
with NHAP. However, among all patients, an etiologic routine use of a broad-spectrum, multidrug regimen is
pathogen was found in 53.5%, with the enterobacter- unnecessary and may lead to overuse of antibiotics. Not
iacea in 19.5%, S. pneumoniae in 12.9%, S. aureus in only do many patients not have infection with MDR
11.6%, P. aeruginosa in 3.6%, and less than 1% had pathogens, but studies in the past, as well as recently,
Acinetobacter infection. The other study [14] random- have demonstrated the efficacy of CAP regimens and of
ized 61 patients with NHAP, not requiring hospital- monotherapy with agents that would not be recom-
ization, to intramuscular therapy, once daily, with mended for HAP patients at risk for MDR pathogens
either cefepime or ceftriaxone. Only six patients had such as levofloxacin, cefepime, and ertapenem [9,14,15].
microbiologic data, but both monotherapies led to In addition, HAP therapy would not adequately treat
similar rates of clinical success: 78% with cefepime some HCAP patients who are infected with atypical
and 66% with ceftriaxone. Although these studies pathogens such as Legionella and Chlamydophila. Finally,
indicate that therapies not directed at MDR pathogens the inclusion of outpatients in the HCAP definition
were effective for patients with NHAP, one study of a requires that some patients get oral therapy in contrast
mixed HCAP population showed that MDR pathogens to intravenous therapy for HAP patients who are already
can be present, and if so, they are often treated in the hospital.
inappropriately [4]. However, in this study, over
90% of the patients were defined as having HCAP On the basis of the available information, we propose an
because of recent hospitalization. algorithm for antibiotic therapy of HCAP that divides

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Healthcare-associated pneumonia: special commentary Brito and Niederman 323

Figure 1 Proposed algorithm for healthcare-associated pneumonia therapy

HCAP is present: From a nursing home, recent hospitalization,

hemodialysis, home infusion therapy

Assess severity of illness (need for mechanical ventilation,

ICU admit) AND

Presence of risk factors for MDR pathogens

(recent antibiotics, recent hospitalization, poor functional status,
immune suppression)

Severe pneumonia

No Yes

0--1 Risks 0 Risks ≥ 1 Risk

≥2 Risks
Treat for common CAP Treat for severe Treat for MDR
pathogens (consider Consider hospital. pneumonia in hospital. pathogens with HAP
oral rx) Quinolone or Treat for MDR Beta-lactam PLUS recommendations. Use
beta-lactam/macrolide pathogens with macrolide or quinolone 3 drugs
HAP therapy

This proposed algorithm, which needs validation, suggests that all patients with HCAP should be identified and then divided on the basis of severity of
illness to guide initial therapy. Patients in each group are then further divided based on whether they have risk factors for drug-resistant pathogens
(MDR pathogens) that include recent antibiotic therapy in the past 6 months, recent hospitalization in the past 3 months, the presence of immune
suppression, and poor functional status as defined by activities of daily living. CAP, community-acquired pneumonia; HAP, hospital-acquired
pneumonia; HCAP, healthcare-associated pneumonia; MDR, multidrug resistant.

patients into four groups based on an assessment of factors who should be considered for hospitalization to
severity of illness (need for mechanical ventilation or receive intravenous therapy directed at MDR pathogens,
ICU admission) and the presence of risk factors for MDR with a two or three-drug HAP regimen. For this latter
pathogens (Fig. 1). The risk factors associated with a high group, this could be achieved with an antipseudomonal
frequency of MDR pathogens in the HCAP studies, in b-lactam (cefepime, imipenem, meropenem, and piper-
addition to severe illness, include: immune suppression, acillin/tazobactam) with an antipseudomonal quinolone
recent hospitalization (within the past 3 months), anti- (ciprofloxacin or high-dose levofloxacin) or aminoglyco-
biotic therapy (within the past 6 months), and poor side, with the addition of either linezolid or vancomycin if
functional status [2,11]. In addition to these risk factors, there is concern about MRSA. If a quinolone is not used
it is important to consider local patterns of microbiology, (because of allergy, intolerance, or recent therapy in
as each hospital setting and nursing home has unique the past 3 months), then addition of a macrolide should
bacteriology, and therapy choices must be modified with be considered, along with the other agents, to provide
such local data in mind [17]. atypical pathogen coverage.

Patients with nonsevere illness include those with zero to Patients with severe illness include those with no other
one risk factor, and they can receive CAP therapy, often additional risk factors for MDR pathogens who should
orally, as an outpatient, with either a quinolone (levo- be treated intravenously, generally in the hospital, with
floxacain or moxifloxacin) or a b-lactam (cefuroxime dual therapy involving a b-lactam (cefotaxime, ceftriax-
orally or ceftriaxone, cefotaxime, or ertapenem intrave- one, or ertapenem) with a macrolide or quinolone (levo-
nously) in combination with a macrolide. The nonsevere floxacin or moxifloxacin). Patients with severe illness
illness group also includes patients with at least two risk and at least one other risk factor for MDR pathogens

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 324 Evidence-based medicine

should receive intravenous therapy with a three-drug ings being correlated to the risk factors. It may be
regimen targeting drug-resistant Gram negatives and necessary and possible to weight each of the risk factors
MRSA, similar to patients with complex nosocomial for drug-resistant pathogens differently, as demonstrated
pneumonia and ventilator-associated pneumonia. This in the one study [18] mentioned above. Validation studies
means using an antipseudomonal b-lactam with either of algorithms for HCAP therapy will be difficult, as
an antipseudomonal quinolone or an aminoglycoside, as demonstrated by the heterogeneous populations that
well as coverage for MRSA with either vancomycin have been studied to date (Table 2). Future studies will
or linezolid. need to be better organized and should include individ-
uals with all forms of HCAP, not just those with NHAP,
One limitation of the proposed algorithm is that each and will need to include outpatients as well as inpatients
risk factor for resistance is weighted similarly. However, and those admitted to both the hospital ward and the
a recent publication, using the data set from the study ICU. Patients with immune suppression should be
by Micek et al. [2], suggested that the presence of included, as should those with recent hospitalization,
multiple HCAP risk factors might relate to the fre- but this definition should probably be restricted to those
quency of MDR pathogens, but that not all risk factors hospitalized in the past 30–90 days and not beyond that
are equivalent. In that study, 639 patients with pneu- time frame. Studies will also need to include patients
monia (including CAP and HCAP) were evaluated, and with and without a specific bacteriologic diagnosis. When
risk factors for drug-resistant organisms were studied in studies are restricted to those with an established etio-
all patients (not just the 431 with HCAP) [18]. The logic diagnosis, they tend to include sicker individuals,
presence of HCAP did not, by itself, predict the pre- many of whom are intubated and mechanically venti-
sence of MDR pathogens, and not all HCAP risk factors lated. In order to study those without a bacteriologic
were equivalent at predicting this risk. In fact, the diagnosis, the success of certain therapies should be
predictors of resistance were given different point related to the presence of risk factors for MDR patho-
scores, with four for recent hospitalization, three for gens. As new therapeutic options become available, they
history of nursing home residence, two for hemodialy- too should be specifically tested in HCAP patients, so
sis, and one for admission to the ICU. The frequency of that data from HAP do not need to be extrapolated to
MDR pathogens was less than 20% for those with less this population.
than three points, 55% for those with three to five
points, and 75% for those with more than five points.
Prior hospitalization was identified as the most import- References
ant risk factor for MDR pathogens, but 93.3% of all the
HCAP patients in this study (and by definition, none of 1 Niederman MS, Craven DE, Bonten MJ, et al. Guidelines for the management
the CAP patients) had recently been hospitalized, so of adults with hospital-acquired, ventilator-associated, and healthcare-
associated pneumonia. Am J Respir Crit Care Med 2005; 171:388 –
the discriminating value of this risk may have been 416.
overly exaggerated. 2 Micek ST, Kollef KE, Reichley RM, et al. Healthcare-associated pneumonia
and community-acquired pneumonia: a single-center experience. Antimicrob
Agents Chemother 2007; 51:3568–3573.
3 Kollef MH, Shorr A, Tabak YP, et al. Epidemiology and outcomes of health-
Conclusion care-associated pneumonia: results from a large US database of culture
HCAP is a heterogeneous disease that includes popu- positive patients. Chest 2005; 128:3854–3862.
lations of patients with varying severity of illness and 4 Zilberberg MD, Shorr AF, Micek ST, et al. Antimicrobial therapy escalation and
hospital mortality among patients with HCAP: a single center experience.
reasons for having had contact with the healthcare Chest 2008; 134:963–968.
environment. Different populations are at different risk 5 Carratalà J, Mykietiuk A, Fernández-Sabé N, et al. Healthcare-associated
for infection with MDR pathogens, and not all patients pneumonia requiring hospital admission: epidemiology, antibiotic therapy, and
clinical outcomes. Arch Intern Med 2007; 167:1393–1399.
need a broad-spectrum, multidrug regimen, similar to
6 Martinez-Moragon E, Garca Ferrer L, Serra Sanchis B, et al. Community
complex nosocomial pneumonia. An algorithm for acquired pneumonia among the elderly: differences between patients living at
therapy that accounts for different subpopulations of home and in nursing homes. Arch Bronconeumol 2004; 40:547–552.
HCAP patients is proposed here, and if followed could 7 Muruyama T, Niederman MS, Kobayashi T, et al. A prospective comparison of
nursing home-acquired pneumonia with hospital-acquired pneumonia in
potentially lead to patients receiving appropriate therapy nonintubated elderly. Respir Med 2008; 102:1287–1295.
without the overuse of antibiotics. 8 Kothe H, Bauer T, Marre R, et al. Outcome of community-acquired pneumonia:
influence of age, residence status and antimicrobial treatment. Eur Respir J
2008; 32:139–146.
These therapy recommendations are based on the best
9 Loeb M, Carusone SC, Goeree R, et al. Effect of a clinical pathway to reduce
available data, which are quite limited, and future vali- hospitalizations in nursing home residents with pneumonia: a randomized
dation of this approach is required. In order for this controlled trial. JAMA 2006; 295:2503–2510.
approach to be validated, patients with a wide range of 10 Lescure FX, Locher G, Eveillard M, et al. Community-acquired infection with
healthcare-associated methicillin-resistant Staphylococcus aureus: the role
risk factors and comorbidities must be enrolled and of home nursing care. Infect Control Hosp Epidemiol 2006; 27:1213–
bacteriologic data collected, with the bacteriologic find- 1218.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Healthcare-associated pneumonia: special commentary Brito and Niederman 325

11 El Solh AA, Pietrantoni C, Bhat A, et al. Indicators of potentially drug-resistant 15 Yakovlev SV, Stratchounski LS, Woods GL, et al. Ertapenem versus cefepime
bacteria in severe nursing home-acquired pneumonia. Clin Infect Dis 2004; for initial empirical treatment of pneumonia acquired in skilled-care facilities or
39:474–480. in hospitals outside the intensive care unit. Eur J Clin Microbiol Infect Dis
2006; 25:633–641.
12 El Solh AA, Akinnusi ME, Pineda LA, et al. Diagnostic yield of quantitative
endotracheal aspirates in patients with severe nursing home-acquired pneu- 16 Mylotte JM, Goodnough S, Gould M. Pneumonia versus aspiration pneumo-
monia. Crit Care 2007; 11:R57. nitis in nursing home residents: prospective application of a clinical algorithm.
J Am Geriatr Soc 2005; 53:755–761.
13 Slinin Y, Foley RN, Collins AJ. Clinical epidemiology of pneumonia in hemo- 17 Beardsley JR, Williamson JC, Johnson JW, et al. Using local microbiologic
dialysis patients: the USRDS waves 1, 3, and 4 study. Kidney Int 2006; data to develop institution-specific guidelines for the treatment of hospital-
70:1135–1141. acquired pneumonia. Chest 2006; 130:787–793.
14 Paladino JA, Eubanks DA, Andelman MH, Schentag JJ. Once-daily cefepime 18 Shorr AF, Zilberberg MD, Micek ST, Kollef MH. Prediction of infection due to
versus ceftriaxone for nursing home-acquired pneumonia. J Am Geriatr Soc antibiotic-resistant bacteria by select risk factors for healthcare-associated
2007; 55:651–657. pneumonia. Arch Intern Med 2008; 168:2205–2210.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.