Bacteremic Pneumonia Due to Staphylococcus aureus: A Comparison of Disease Caused by Methicillin-Resistant and Methicillin-Susceptible Organisms
Carmen Gonza lez
Miguel Gonza lez
Juan J. Picazo
Materials and Methods
Department of Biostatistics, Universidad de Extremadura
Received 13 January 1999; revised 17 June 1999. Salinero de los Santos
1 Esc Dcha 37 A, 06011 Badajoz
Department of Medical Specialties, Universidad Europea de Madrid
Department of Medical Microbiology, Hospital Universitario San Carlos
We performed a prospective study of all patients with bacteremic pneumonia due to Staphylococcus aureus over a period of 6 years during an outbreak of methicillin-resistant S. aureus (MRSA). Patients with bacteremic pneumonia due to MRSA (32 cases) or methicillin-susceptible S. aureus (MSSA; 54 cases) were compared. The patients with MRSA pneumonia were older and were more likely than those with MSSA pneumonia to have predisposing factors for acquisition of the infection. There were no differences in clinical findings, radiological pattern, or complications in clinical evolution among patients with MRSA and MSSA pneumonia. Mortality was significantly higher among MSSA-infected patients treated with vancomycin than among those treated with cloxacillin (47% vs. none; P ! .01). Multivariate analysis (stepwise logistic regression method) showed a relationship between mortality and the following variables: septic shock (odds ratio [OR], 61), vancomycin treatment (OR, 14), and respiratory distress (OR, 8).
Pneumonia due to Staphylococcus aureus constitutes 1%10%
of cases of community-acquired pneumonia  and up to
16% of cases of nosocomial pneumonia . Its clinical
manifestations do not appear to differ from those of pneumonias
due to other bacterial etiologies , and its radiological
pattern is heterogeneous [5, 810], in many cases associated with
cavitation [5, 8]. The mortality is high [5, 7, 1113].
Numerous epidemic outbreaks of methicillin-resistant S.
aureus (MRSA) infection have been reported in hospitals in
Europe, the United States, and Australia over the past 3 decades
. In 1990 there began an enormous outbreak of
infections due to this microorganism in our hospital, and outbreaks
began at the same time in other Spanish hospitals .
MRSA pneumonia was the second most frequent infection
after surgical wound infection in our center and was a major
factor associated with mortality . The high frequency of this
infection during MRSA epidemics has also been described in
other studies [18, 23]. However, there exist few studies on
pneumonia due to this microorganism [12, 25] or comparative
analyses indicating the differences between pneumonia due to
MRSA and to methicillin-susceptible S. aureus (MSSA) [13,
26]. The difference in virulence of these 2 microorganisms also
is not known. We therefore performed a prospective study of
Hospital. The Hospital Universitario San Carlos is a tertiary
care center with 1479 beds, serving a population of 569,307
inhabitants northwest of the Madrid metropolitan area. During the
study period (1 January 1990 to 31 December 1995), 228,177
patients were hospitalized in this center. The hospital has all the
medical, surgical, and neonatal specialities, as well as intensive care
and postsurgical recovery units. Kidney and bone marrow
transplants are performed. There is no burn unit. There is a 16-bed unit
for the admission of HIV-infected patients and a day hospital for
their outpatient care.
Microbiological methods. Samples for microbiological
evaluation were taken by bronchial aspiration, bronchoalveolar lavage,
and protected specimen brush and for hemoculture. The bronchial
aspirates were collected in sterile receptacles. Sample quality was
evaluated according to the criteria of Murray and Washington ,
to avoid samples with high levels of oropharyngeal contamination.
We consider the finding of S. aureus in pure culture of 2 samples
of bronchial aspirate from the same patient on consecutive days
to be significant for the diagnosis of pneumonia. Samples were
obtained by protected specimen brush according to the technique
described by Wimberley , and samples were immediately seeded
on reaching the laboratory. S. aureus colony counts of >103 cfu/
mL were considered to be significant. Bronchoalveolar lavage was
also done by means of bronchoscopy. Quantitative culture was
performed on arrival of the samples at the laboratory. S. aureus
at >105 cfu/mL was considered significant. Hemocultures were
processed by use of the BACTEC-NR 660 system until April 1992
and by use of the BACTEC-NR 860 system after that time (both
Becton Dickinson, Sparks, MD).
S. aureus was identified by the presence on blood culture of
characteristic colonies of gram-positive cocci that were
catalaseand coagulase-positive in plasma after 24 h of incubation.
Antibiotic sensitivity was determined by the Mueller-Hinton agar
diffusion technique, following the recommendations of the National
Committee for Clinical Laboratory Standards.
Study population. Patients with bacteremic pneumonia due to
S. aureus during 1 January 1990 to 31 December 1995 were studied
prospectively. The following data were recorded for each patient:
age, sex, hospitalization time, length of hospitalization before
acquisition of the infection, area of hospitalization, factors
predisposing to infection, underlying diseases, and antibiotics used in
treating the pneumonia, together with the administration route,
dose, and duration of therapy. These patients were followed until
their recovery or death. In the case of death, the cause (infection
or underlying disease) was recorded. For each of the patients with
a blood or respiratory tract sample that yielded S. aureus on culture,
all the information available from the microbiology service
referring to the isolated microorganism and its antibiotic sensitivity was
Study design. The study included those patients diagnosed as
having bacteremic pneumonia due to MRSA or MSSA between 1
January 1990 and 31 December 1995. Epidemiological and clinical
characteristics, the clinical evolution, and mortality of the 2 groups
Definition of terms. A diagnosis of pneumonia was made when
there existed symptoms of infection of the lower respiratory tract
(e.g., cough, purulent expectoration, chest pain) and pulmonary
infiltrates on the chest radiograph not attributable to other causes,
coinciding with the isolation of S. aureus as the only pathogen in
2 samples of bronchial aspirate taken on consecutive days or in 1
or more of the following samples: protected specimen brush,
bronchoalveolar lavage, transthoracic puncture, and/or hemoculture. S.
aureus bacteremia was defined as 1 or more hemocultures that were
positive for this microorganism. Significant bacteremia was defined
as the isolation of S. aureus from at least 1 hemoculture,
accompanied by symptoms of clinical sepsis unexplained by other causes.
Bacteremic pneumonia was defined as the diagnosis of both
pneumonia and significant bacteremia in a patient, by use of the
definitions established above. Following the criteria of Finegold ,
aspiration pneumonia was defined as cases in which there was a
clear episode of aspiration of particulate matter, foreign body or
fluid (normally food), or gastric content in the 72 h prior to the
diagnosis of pneumonia.
Nosocomial infection was defined as those cases in which the
clinical manifestations began 72 h after hospital admission and the
sample positive for S. aureus was taken during this period or,
alternatively, before this period if the patient had been admitted in
the 2 previous weeks. In the cases in which the positive sample
(respiratory and/or hemoculture) was obtained before the first 72
h after admission and the patient had not had an admission during
the previous 2 weeks, the infection was considered to be community
acquired. The presence of neutropenia was defined as a blood
polymorphonuclear neutrophil count of <500/mL.
There was considered to be prior chemotherapy when there had
been oncological chemotherapy in the month preceding the
diagnosis of the infection and/or steroid treatment in patients who
received a dose of at least 10 mg of prednisone/day (or equivalent)
over at least 48 h in the 2 weeks preceding diagnosis of the infection.
Underlying diseases were classified according to the criteria of
McCabe and Jackson  into rapidly fatal, eventually fatal, or
nonfatal. Clinical status during the pneumonia episode was
classified according to Winston et al.  into critical, bad, poor, or
An alveolar radiological pattern was defined as the presence of
alveolar lesions that appeared as an increase in density of alveoli.
This pattern presented with poorly defined and fuzzy margins,
tending to coalesce to form a lobar or segmentary distribution or, in
the case of massive diffuse affectation, a butterfly wing
distribution. An interstitial radiological pattern was defined as the
appearance of pulmonary parenchymal lesions in the form of lines,
small nodules, or rounded cysts of !1 cm in diameter or a mixture
of these. A mixed pattern was defined as the appearance of both
alveolar and pulmonary parenchymal lesions. Nodular lung
condensations (single or, more frequently, multiple) with sharply
defined edges, with or without cavitation, were considered to be septic
emboli. Radiologically, lung cavitation was considered to be the
presence of a sharp zone of parenchymal loss, delimited by a wall
and filled with liquid or air. Pleural effusion was the filling of the
pleural cavity with liquid, and empyema was the existence in said
cavity of a clearly delimited mass directed toward the lung
containing a hydro-aerial level.
Shock was defined as a reduction in systolic arterial pressure to
!90 mm Hg or in diastolic pressure to !60 mm Hg, accompanied
by clinical signs of hypoperfusion, such as oliguria or peripheral
vasoconstriction. In patients with previous hypertension, it was
defined as a 20% decline in the prior values. Acute renal
insufficiency was defined as the sudden appearance of oliguria with an
increase in urea and serum creatinine. Respiratory distress was
defined as the appearance or aggravation of respiratory
insufficiency due to a diffuse disorder of alveolar-capillary permeability,
manifested as an increase in pulmonary capillary resistance .
Septic hepatopathy was defined as an increase in transaminase and
alkaline phosphatase levels above twice the normal values (or above
twice the prior value for patients with hepatic disease) after
exclusion of other causes of increases in these parameters. The presence
of CNS disease was defined as the existence of reduced awareness
or the appearance of focal neurological signs. There was considered
to be disseminated intravascular coagulation when there was
thrombopenia of <80,000 platelets/mL in the presence of fibrin
Antibiotic treatment. The antibiotic treatment administered
and its duration was determined by the clinician. We considered
appropriate antibiotic treatment to be the administration of drugs
that are active in vitro against the isolated strain of S. aureus, that
had recognized clinical efficacy, and that were administered in
therapeutic doses. The efficacy of the treatment was considered to be
assessable when administration was over a period of at least 48 h,
the patient did not die as a consequence of the underlying disease
before the completion of the treatment, and the underlying clinical
situation allowed the response to the antibiotic treatment to be
Table 1. Characteristics and underlying diseases of patients with
pneumonia due to Staphylococcus aureus.
NOTE. Data are no. (%) or mean 5 SD. ICU, intensive care unit; MRSA,
methicillin-resistant S. aureus; MSSA, methicillin-susceptible S. aureus.
a Statistically significant.
b Adapted from .
Clinical evolution. Clinical evolution was classified as cure
when the clinical manifestations disappeared and the cultures
became negative and as failure when there was persistence of the
clinical manifestations, persistence of positive cultures, or evidence
of death directly due to the staphylococcal infection. Death was
considered to be due to the staphylococcal infection when at least
1 of the following criteria was satisfied: a culture positive for S.
aureus from a sample obtained at the time of death (hemoculture
or representative sample from the lower respiratory tract),
persistent respiratory symptoms, radiological evidence of the persistence
of the pneumonia, or occurrence of death without any explanation
within the first 7 days after the diagnosis of staphylococcal
pneumonia. Deaths that did not satisfy one of these criteria were
considered to be a consequence of the patients underlying disease.
Follow-up. The follow-up period was from the diagnosis of
pneumonia (by a result positive for S. aureus on culture of a
respiratory sample or hemoculture) until cure or death.
Statistical analysis. The analysis of variables was done by use
of the statistical package BMDP V7.0 (BMDP, West Los Angeles,
CA) for UNIX. The Mann-Whitney test was used for the
comparison of the quantitative data. Qualitative variables were
associated by means of the x2 test and, when the conditions of validity
of this test were not fulfilled, by means of Fishers exact test.
Statistical significance was considered to be P ! .05.
A stepwise logistic regression model was defined, to evaluate the
factors associated with mortality due to staphylococcal pneumonia.
The magnitude of the effect was measured from the corresponding
ORs and from the 95% CIs for the same.
General data. During the study period, S. aureus was the
cause of 134 cases of bacteremic pneumonia. A total of 36 cases
(1 episode/patient) were caused by MRSA (26.9% of the total
number of bacteremic pneumonia cases), and 98 cases (73.1%)
were due to MSSA. For the comparative analysis between the
MRSA and MSSA cases of pneumonia, we excluded those
patients who were parenteral drug users (48, 44 with MSSA
and 4 with MRSA infections), since nearly all of them had
community-acquired MSSA pneumonia with characteristics
that were well differentiated from those of the rest of the study
population (pulmonary septic emboli from tricuspid
endocarditis with unpronounced respiratory symptoms, generally few
complications, and a good prognosis). We analyzed 86 patients:
32 (37.2%) with MRSA and 54 (62.8%) with MSSA infections.
The epidemic of MRSA infections began in our center at the
beginning of 1990, and the percentage of MRSA isolates
identified in the microbiology laboratory increased to 56% of the
total S. aureus cases in that year. In October 1990, strict control
measures were set up, and the frequency of MRSA in the
following year varied from 41%54%. A decline in frequency was
noticeable by the end of 1992, and at the end of the study
period (1995) the percentage of MRSA as a total of S. aureus
cases was 20%.
Clinical data. Table 1 shows the general characteristics and
underlying diseases of the 2 groups of patients. The cases of
MRSA pneumonia were most often nosocomial (P ! .01 ), and
Table 2. Predisposing factors among patients with pneumonia due
to Staphylococcus aureus.
NOTE. Data are no. (%). ICU, intensive care unit; MRSA,
methicillinresistant S. aureus; MSSA, methicillin-susceptible S. aureus.
a Statistically significant. b Adapted from .
Deaths due to infection
Deaths in the first 48 h
NOTE. Data are no. (%). MRSA, methicillin-resistant S. aureus; MSSA,
methicillin-susceptible S. aureus.
the patients with pneumonia due to this microorganism were
more elderly than the MSSA-infected patients (P = .01).
Also, the mean number of days of hospitalization before
acquisition of the pneumonia was significantly greater (P !
.01) for those with MRSA infection. As can be seen in table
1, there were no differences with regard to the areas of
hospitalization between the 2 groups of patients.
With respect to underlying diseases, it is noteworthy that the
mean number of underlying diseases was significantly greater
(P = .01) among those with MRSA pneumonia (2.31 5 1.3)
than among those with MSSA pneumonia (1.6 5 1.15). The
gravity of these underlying diseases was measured according to
the criteria of McCabe and Jackson . Of the patients with
MRSA pneumonia, 13 (40.6%) were in a rapidly fatal situation,
compared with 13 (24.1%) with MSSA pneumonia, but the
difference was not significant (P = .1). Lung diseases and
vascular diseases were significantly more frequent among the
patients with MRSA pneumonia than among those with MSSA
pneumonia (34.3% vs. 16.6% [P = .05] and 15.6% vs. 0% [P !
The factors predisposing to infection are listed in table 2. Of
the patients with bacteremic pneumonia due to MRSA, 37.5%
had undergone antibiotic treatment before the pneumonia
episode, compared with 7.5% of patients with bacteremic
pneumonia due to MSSA (P ! .01). The frequencies of previous
pneumonia (P = .04), urinary catheters (P ! .01), and central
catheters (P ! .01) were significantly higher among patients with
MRSA bacteremic pneumonia.
The clinical status ratings of the patients, according to the
criteria of Winston et al. , were very similar in the 2 groups.
In the MRSA group, the radiological pattern was alveolar in
all patients, as it was in most (47 patients [87%]) of the MSSA
group. Also, in this latter group, there were 4 patients (7.4%)
with a nodular pattern, 1 (1.9%) with a mixed pattern, and 2
(3.7%) in which the radiological pattern could not be
determined. One (3.1%) of the patients with pneumonia due to
MRSA developed cavitation, as did 1 (1.9%) of the patients
with pneumonia due to MSSA (P = 1). The differences in the
percentages of pleural effusion that developed in the 2 groups
were also not significant (6.3% MRSA pneumonia vs. 16.7%
MSSA pneumonia; P = .19). The involvement was multilobular
in 46.9% of the MRSA group and 53.7% of the MSSA group
(P = 1.00), and there were no differences between the 2 groups
with respect to the presence of diffuse bilateral involvement: 7
patients (21.9%) in the MRSA group and 16 patients (29.6%)
in the MSSA group (P = .27).
Treatment and evolution. Of the 32 patients with MRSA
pneumonia, 22 received appropriate antibiotic treatment, and
the evolution of their infection could be evaluated. Of these,
20 were treated with vancomycin (2 of them received some doses
of a third-generation cephalosporin during the course of the
treatment). Two patients were treated with teicoplanin
(associated with co-trimoxazole, and with clindamycin in the other).
Of the remaining 10 patients, 8 died within the first 48 h after
diagnosis (table 3), 1 died as a consequence of the underlying
disease, and 1 had received nonactive antibiotics.
Of the total of 54 patients with MSSA pneumonia, 41
received appropriate antibiotic treatment, and the evolution of
their infection could be evaluated. Ten patients were treated
with cloxacillin, which was associated with aminoglycosides in
3 patients and with third-generation cephalosporins in another
3. Vancomycin was most often used, with 17 patients receiving
it (2 received a dose of a third-generation cephalosporin, and
another received imipenem during the course of the treatment).
Two patients were initially treated with vancomycin and, when
the results of the sensitivity of the microorganism were known,
had their treatment changed to cloxacillin; in contrast, there
were 4 patients whose initial treatment was cloxacillin, which
was changed to vancomycin during the course of the infection.
Three patients received teicoplanin (associated with a
thirdgeneration cephalosporin in 1 patient); other drugs with S.
aureus activity (mainly b-lactam antibiotics) were given as
treatment to 5 patients: cephalothin plus aminoglycosides,
cefotaxime plus erythromycin, and cefotaxime plus clindamycin. Of
the remaining 13 patients, 8 died in the first 48 h after diagnosis
of the infection (table 3), 4 died as a consequence of the
underlying disease, and 1 was moved to another hospital.
The mortality associated with the infection was high for both
groups of patients. Of the 32 patients with MRSA pneumonia,
18 (56.3%) died as a direct result of the infection, and, of the
54 patients with MSSA pneumonia, 22 (40.7%) died (P = .17;
table 3). The percentage of complications that developed was
similar for the 2 groups (table 4).
Of the 22 patients with MRSA pneumonia who were
correctly treated and whose evolution could be evaluated, 11 (50%)
died, and of the 41 with MSSA pneumonia, 14 (34%) died
(P = .34). The mortality associated with the infection in patients
correctly treated with vancomycin was very high in both the
MRSA and the MSSA groups (50% vs. 47%; not significant).
Table 4. Clinical evolution among patients with pneumonia due to
NOTE. Data are no. (%). MRSA, methicillin-resistant S. aureus; MSSA,
methicillin-susceptible S. aureus.
Table 5. Antibiotic treatment among patients with pneumonia due
to Staphylococcus aureus.
No. treated (%)
No. who died (%)
NOTE. MRSA, methicillin-resistant S. aureus; MSSA,
methicillin-susceptible S. aureus.
a P ! .01.
The infection-associated mortality among patients with MSSA
pneumonia was significantly greater for those treated with
vancomycin than for those treated with cloxacillin (47% vs. 0%;
P ! .01). These data are given in table 5.
Of the total of 63 patients who received appropriate antibiotic
treatment and whose clinical evolution could be evaluated, a
number of factors were associated with mortality (age, number
of underlying diseases, urinary catheter, neurological
alterations, septic shock, respiratory distress, acute renal
insufficiency, and having undergone treatment with vancomycin).
These factors were included in a stepwise logistic regression
model, which predicted adequately an average of 85.7% of the
observed data. The model showed that vancomycin treatment
was independently associated with mortality from the infection
and that the risk of death was 14 times greater (OR, 14.5) for
vancomycin-treated patients than for those who had received
other treatments. Other variables that were found to be
independently associated with mortality were septic shock (OR,
61.5) and respiratory distress (OR, 8.3) (table 6). The other
variables introduced into the model as potentially important
for prognosis (age, number of underlying diseases, neurological
alterations, presence of a urinary catheter, and acute renal
insufficiency) produced no differences in the ORs of the 3
previously analyzed variables.
Considering all the possible interactions among the 3
variablesseptic shock, vancomycin treatment, and respiratory
distresswe observed, first, independently of the presence of
either of the other 2 variables, that the presence of shock
increased the likelihood of death 47-fold (OR, 47.29). Second, in
the absence of respiratory distress, vancomycin treatment did
not increase mortality. In the presence of distress, however,
treatment with vancomycin increased the likelihood of death
38-fold (OR, 38.46) (table 7).
During the last 3 decades, the epidemiology and
microbiology of epidemic outbreaks of MRSA infection have been
analyzed in a number of studies . There have been fewer
publications, however, describing the characteristics of patients
with pneumonia due to MRSA [12, 13, 25, 26].
The epidemic of MRSA infection in our hospital that began
in 1990 reached large proportions and is one of the most
extensive reported in the literature. This circumstance gave us the
opportunity of studying a good number of patients with
pneumonia due to MRSA. Of the total cases of bacteremic
pneumonia studied, 27% were due to MRSA. The frequency varied
from 40% to 50% when those of nosocomial origin were
considered, depending on whether the pneumonia was primary or
In general, MRSA pneumonia, as with other infections
caused by this organism, is usually of nosocomial origin [7, 12,
33]. There exist no broad studies clarifying its incidence in the
community . Some community-acquired cases have been
found, however, in series of staphylococcal pneumonia  and
among cases related to infections in nursing homes . Series
of community-acquired MRSA bacteremias have also been
described in patients who use drugs parenterally [36, 37]. We had
6 cases of community-acquired pneumonia due to MRSA: 2
elderly patients with aspiration pneumonia and 4 drug-using
patients with septic emboli pneumonia.
The factors associated with MRSA pneumonia in our
patients were advanced age, prolonged hospitalization, gravity of
the underlying diseases, especially lung diseases (34.3%), prior
antibiotic therapy, and surgery or other invasive maneuvers
(vesicular probe, central catheters, and endotracheal
intubation). There were no differences in clinical findings, radiological
pattern, or complications in clinical evolution among patients
with MRSA pneumonia, compared with patients with MSSA
pneumonia. Our MRSA-infected patients had a high mortality
(56.3%). In a study published in 1994, Iwahara et al. 
analyzed 32 patients with MRSA pneumonia and found these
same risk factors for the acquisition of the infection, with the
exception of the lung diseases. They did not study MSSA
pneumonia. The infection-associated mortality was lower (38%). The
greater mortality in our study may be attributed to the high
percentage of patients with grave lung disease and to the fact
that all the pneumonia cases analyzed were bacteremic.
Rello et al. , in a comparative study of pneumonia due
to MRSA and MSSA in patients in an intensive care unit, found
the same epidemiological differences as we did and also
described 2 important consequences. First, MRSA pneumonia
produced a significantly greater frequency of bacteremia (36.4%
vs. 10.5%; RR, 3.4) and septic shock (27.3% vs. 7.9%; RR, 3.4).
Second, the infection-associated mortality in their study was
greater among the patients with MRSA pneumonia (54.5% vs.
2.6%; RR, 20.7). In the design of our study, all our patients
had bacteremic pneumonia, which we then split into MRSA
and MSSA bacteremic pneumonia. Therefore, the first
Table 7. Factors associated with mortality among patients with
pneumonia due to Staphylococcus aureus when all possible interactions
between shock, vancomycin treatment, and respiratory distress were
in presence of distress
quence of the study of Rello et al.  could not be tested, and
we think that further studies will have to be done to allow this
question to be clarified. Our mortality figures also differed from
those of Rello et al. : Although we found a greater mortality
among those with pneumonia due to MRSA than among those
with pneumonia due to MSSA, the difference was not
significant (56.3% vs. 40.7%; P = .17). The high mortality among our
patients with MSSA pneumonia was probably influenced by
the high proportion of patients who received vancomycin as
treatment, a factor that, via the logistic regression model in our
design, we found to be directly related to an increase in
The mortality of our MRSA patients who were correctly
treated with vancomycin was very high (50%). There are no
reports in the literature that allow us to contrast this effect.
Recently, Georges et al.  demonstrated a low level of
vancomycin in the lungs at 24 h after initiation of treatment in 10
patients with MRSA pneumonia. Further studies will be
required, however, to confirm the clinical relevance of this finding.
In the group of correctly treated patients with MSSA
pneumonia, we found that the infection-associated mortality was
significantly higher among vancomycin-treated patients (47%)
than among those treated with cloxacillin (0%; P ! .01 ). Recent
studies  have shown suboptimal results in the treatment
of bacteremia and endocarditis with vancomycin; however, this
is the first study that shows these suboptimal results in patients
with pneumonia due to S. aureus. The reason for these findings
has not been clarified. In some in vitro studies, the bactericidal
action of vancomycin has been shown to be less and slower
than that of cloxacillin . Other authors  attribute this
poorer effect of vancomycin to its use for treatment of patients
with more-grave disease and worse prognosis. In our work, we
determined the influence of vancomycin treatment on the
mortality among patients with staphylococcal pneumonia. We
therefore believe it necessary to carry out future investigations,
to perfect the treatment of this infection and to control a series
of now known epidemiological factors that may lead to
reducing the mortality.
1. Fang GD, Fine M, Orloff J, et al. New emerging etiologies for
communityacquired pneumonia with implications for therapy: a prospective
multicenter study of 359 cases. Medicine 1990; 69:30716.