Download Multicenter, double-blind, placebo-controlled study of the use of

Multicenter, double-blind, placebo-controlled study of the use of
filgrastim in patients hospitalized with pneumonia and severe sepsis*
Richard K. Root, MD; Robert F. Lodato, MD, PhD; Ward Patrick, MD; John Francis Cade, MD, PhD;
Nick Fotheringham, PhD; Steven Milwee, PharmD; Jean-Louis Vincent, MD, PhD; Antonio Torres, MD;
Jordi Rello, MD; Steve Nelson, MD; for the Pneumonia Sepsis Study Group
Objective: To determine the safety and efficacy of filgrastim
(r-metHuG-CSF) in combination with intravenous antibiotics to
reduce the rate of mortality in patients with pneumonia and
sepsis.
Design: This study was multicenter, double-blind, and randomized.
Setting: Intensive care units
Patients: Adult patients with bacterial pneumonia, either acquired or nosocomial, as confirmed by chest radiograph and
positive culture or Gram-negative stain, and severe sepsis, defined as sepsis-induced hypotension or organ dysfunction.
Interventions: Standard antibiotic therapy with or without filgrastim (300 ␮g/day) or placebo administered as a 30-min intravenous infusion. The study drug was started within 24 hrs of
enrollment and was continued for 5 days or until the white blood
cell count reached >75.0 ⴛ 109 cells/L.
Measurements and Main Results: The primary end point was
the occurrence of mortality through day 29; secondary end points
included occurrence of subsequent organ dysfunction, time to
discharge from intensive care unit, number of days on mechanical
A
study of the epidemiology of
severe sepsis in the United
States determined the national
prevalence to be 751,000 cases
for the calendar year 1995 with a crude
mortality of 28.6% or 215,000 cases (1).
The annual increase in severe sepsis cases
*See also p. 635.
University of Washington, Seattle, WA (RKR), Amgen, Thousand Oaks, CA (SM, NF), The University of
Texas Health Sciences Center, Houston, TX (RFL), Hopital Erasme, Brussels, Belgium (J-LV), Queen Elizabeth II Health Sciences Center, Halifax, Nova Scotia,
Canada (WP), Hospital Clinical I Provincial, Barcelona,
Spain (AT), The Royal Melbourne Hospital, Parkville,
Australia (JFC), Hospital Universitari Joan XXIII, Tarragona, Spain (JR), Louisiana State University Health
Sciences Center, New Orleans, LA (SN).
Supported by Amgen, Thousand Oaks, CA.
Address requests for reprints to: Steve Nelson,
MD; John H. Seabury Professor of Medicine, Pulmonary/Critical Care Medicine, Louisiana State University
Medical Center, Suite 3205, MEB, 1901 Perdido St,
New Orleans, LA 70112. E-mail: [email protected]
Copyright © 2003 by Lippincott Williams & Wilkins
DOI: 10.1097/01.CCM.0000048629.32625.5D
Crit Care Med 2003 Vol. 31, No. 2
ventilatory support, and time to death. Study-related observations
were recorded through day 10 and included vital signs, onset of
organ dysfunction, clinical laboratory variables, and adverse
events. Filgrastim increased the white blood cell count to a
median peak of 31.7 ⴛ 109 cells/L from a baseline of 12.3 ⴛ 109
cells/L. The two groups were well matched and did not differ
significantly with regard to severe adverse events, time to death,
occurrence of end-organ dysfunction, days of intensive care unit
hospitalization, or days on mechanical ventilatory support. Mortality was low in both treatment groups; the mortality rate in
patients with adult respiratory distress syndrome was similar
between the two groups.
Conclusions: The addition of filgrastim to the antibiotic and
supportive care treatment of patients with pneumonia complicated by severe sepsis appeared to be safe, but not efficacious in
reducing mortality rates or complications from this infection. (Crit
Care Med 2003; 31:367–373)
KEY WORDS: pneumonia; sepsis; clinical trials; phase 3; filgrastim; mortality; multiple organ failure
was projected to be 1.5%, with the burgeoning elderly population accounting
for both the highest incidence and mortality rates. Despite improvements in antimicrobial therapy and supportive care,
these data indicate that improved approaches for the management of severe
sepsis are a major unmet medical need
(2).
The lungs constitute the site of origin
in at least 50% of the patients with severe
sepsis (1, 3). With 6 million cases,
500,000 hospitalizations, and 50,000
deaths from community-acquired pneumonia occurring annually (4), new strategies are needed beyond the appropriate
use of antibiotics and proper respiratory
and supportive care to improve outcomes
in patients with severe cases of pneumonia. Preclinical studies demonstrated that
filgrastim (r-metHuG-CSF) can enhance
both neutrophil production and function,
and improve survival in a variety of animal models of pneumonia (5). In a large
(756 patients) placebo-controlled trial,
filgrastim given as an adjunct to antibiotics in the treatment of severe community-acquired pneumonia promptly increased circulating neutrophil counts
and was well tolerated (6). Furthermore,
in a large subgroup of the patients who
had multilobar disease, filgrastim treatment accelerated radiologic resolution,
reduced complications, and shortened
hospitalization in the intensive care unit
(6). In various experimental models of
sepsis, filgrastim-treated animals have
lower mortality rates associated with
greater clearance of bacteria and endotoxin and a reduction in cytokine levels
compared with placebo controls. Results
of a small phase 1/2 pneumonia sepsis
study in a severely ill population (Acute
Physiology and Chronic Health Evaluation II scores ⬎25) demonstrated the
ability of filgrastim to increase absolute
neutrophil counts; filgrastim-treated patients had improved resolution of shock
367
and a lower occurrence of mortality (7).
Filgrastim was well tolerated and, specifically, acute respiratory distress syndrome rates were not different between
placebo and filgrastim-treated groups.
Based on these encouraging preclinical
and clinical data, it was hypothesized that
the addition of filgrastim to a regimen of
intravenous antibiotics would reduce the
rate of mortality in patients with pneumonia and sepsis.
MATERIALS AND METHODS
Study Design. The institutional review
boards of all participating institutions approved the study design, and all patients or
legal guardians gave written informed consent
before any study-specific tests were done.
The study design was multicenter, doubleblind, randomized, and placebo-controlled.
Patients who met the eligibility requirements
were randomized to receive filgrastim or placebo in a 1:1 ratio using a predetermined
randomization list. A separate randomization
list was generated for each center. Study drug
was initiated as soon as possible (always within
24 hrs of meeting study entry criteria) and was
continued for 5 days or until the white blood
cell count was ⬎75.0 ⫻ 109 cells/L. Studyrelated observations, consisting of vital signs,
onset of organ dysfunction, laboratory values,
and adverse events, continued through day 10.
Patient survival on day 29 was determined.
Patients. Adult patients (⬎18 yrs old) hospitalized with either community-acquired or
nosocomial bacterial pneumonia were eligible
for study entry if they met specific inclusion
criteria. These criteria included presence of
infiltrates on chest radiograph compatible
with a diagnosis of bacterial pneumonia, confirmed by a positive culture of the blood, protected specimen brush, or bronchoalveolar lavage (for patients on ventilatory support ⱖ3
days). Patients had to have fever (temperature
ⱖ38°C), tachypnea (⬎20 breaths/minute or
need for mechanical ventilatory support), and
tachycardia (heart rate ⬎90 beats/minute). In
addition, patients had to have a diagnosis of
pneumonia-induced severe sepsis by at least
one of the following organ dysfunctions: sepsis-induced hypotension, acute respiratory
distress syndrome, disseminated intravascular
coagulation, acute oliguria, or lactic acidosis
as defined in the protocol.
Patients were not eligible for participation
if they were pregnant or breastfeeding; had a
life expectancy (unrelated to the acute infection) of ⬍72 hrs, cardiogenic shock as the
primary acute condition, uncontrolled hemorrhage, or the presence of full-thickness thermal or chemical burns (⬎20% of body surface); had a white blood cell count ⬎40 ⫻ 109
cells/L, had history of New York Heart Association class IV heart failure; had do-notresuscitate orders, or had known hypersensitivity to Escherichia coli-derived products.
368
Treatment. Filgrastim (300 ␮g/day) or placebo (vehicle in an identical vial) (Amgen,
Thousand Oaks, CA) were administered as a
30-min intravenous infusion for 5 days, until
the white blood cell count was ⬎75 ⫻ 109/L,
or until the day intravenous antibiotics were
discontinued, whichever came first. The study
drug was administered as soon as possible and
within 24 hrs of meeting criteria for severe
sepsis and within 4 hrs of randomization. Subsequent doses were given at 24-hr intervals
after the first dose and within 5 hrs after the
results of a white blood cell count were
known.
Supportive Care and Monitoring. Full supportive care was given as required. Concomitant medications, including antibiotics, were
administered as medically indicated by the
treating clinicians. Baseline assessments included clinical evaluation (vital signs, Acute
Physiology and Chronic Health Evaluation II
score, electrocardiogram, chest radiograph,
arterial blood gases, and sputum and blood
cultures). During the course of the study, vital
signs and complete blood counts were recorded daily, and patients were monitored for
organ dysfunction, adverse events, concomitant medications, blood transfusions, and procedures. Chest radiographs, arterial blood
gases, and other blood panels were done as
clinically indicted.
Study End Points and Sample Size. The
primary efficacy end point was the occurrence
of mortality through day 29. Secondary efficacy end points included occurrence of subsequent end-organ dysfunction through day 10,
days on mechanical ventilatory support
through day 29, time to death, and time to an
intensive care unit discharge. The safety end
points were the occurrence of adverse events
through day 10, changes from baseline in laboratory parameters, and mortality through
day 29.
A sample size of 700 patients was calculated to provide 80% power to detect a reduction in 28-day mortality from 35% in the placebo group to 25% in the filgrastim-treated
group. The estimated placebo rate was based
on the observed mortality rates in the sample
populations with severe sepsis enrolled in previous Amgen-sponsored studies as well as the
published literature.
Statistical Methods. The proportions of
deaths occurring on or before day 29 and the
proportion of patients with new end-organ
dysfunction on or before day 10 were compared using the Cochran-Mantel-Haenszel test
with stratification by center. A center-pooling
algorithm was devised in which centers were
pooled to ensure that each pooled center had
at least one patient on each treatment and at
least one experiencing each outcome.
Kaplan-Meier estimates of the distributions of survival time and time to intensive
care unit discharge were compared between
treatment groups with the log-rank test, and
days on mechanical ventilatory support were
compared with the Wilcoxon’s rank-sum test.
Adverse events were tabulated by body system,
severity, and relationship to study drug. No
interim analysis was planned, and none was
performed.
RESULTS
Patients
From April 1996 through December
1998, 701 patients from 96 hospital centers in the United States, Canada, Australia, and Europe were enrolled in this
study and randomized to receive filgrastim (n ⫽ 348) or placebo (n ⫽ 353).
Forty-four patients (6.3% of all patients)
did not complete the study for reasons
other than death.
The two study groups were balanced
in terms of age, sex, race, and baseline
comorbid medical disorders, including
vital signs and blood gases (Table 1). All
patients who participated in this study
presented with either community-acquired pneumonia (80%) or nosocomial
pneumonia (20%); the distribution was
approximately equal in both groups.
A blood culture was performed at
baseline for 672 patients (96%) (Table 2),
and a respiratory secretion culture was
performed at baseline for 625 patients
(89%) (Table 3). Twenty-nine percent of
the patients were bacteremic, with Streptococcus pneumoniae and Staphylococcus aureus the most frequently found
pathogens.
Drug Administration and
Responses
A total of 699 patients received at least
one dose of study drug and were evaluable for safety. Two patients died before
receiving study drug (one in each group).
Eighty-four percent of all patients received the complete 5-day course of study
drug. The median highest white blood
cell count (i.e., the median of the highest
individual patient values) was 19.3 ⫻ 109
cells/L in the placebo group and 31.7 ⫻
109 cells/L in the filgrastim group (Fig.
1). The median time to each peak white
blood cell count was 9 days in the placebo
group and 4 days in the filgrastim group.
Once filgrastim therapy was discontinued, the white blood cell count returned
to baseline, usually within 3 days.
The most frequently prescribed antibiotics were third-generation cephalosporins (64% to 66%), followed by macrolides (53% to 56%), and aminoglycosides
(40% to 44%). This usage is consistent
Crit Care Med 2003 Vol. 31, No. 2
Table 1. Summary of key demographic and baseline characteristics and risk factors in patients with
pneumonia and severe sepsis
Placebo
(n ⫽ 353)
Filgrastim
(n ⫽ 348)
60.0 (16.4)
19–93
58.9 (17.1)
16–93
with the broad-spectrum coverage administered to patients with severe pneumonia.
Study End Points
Age, yrs, mean (SD)
Age range, yrs
Sex, n (%)
Men
Women
Risk factors, n (%)
Anemia
Diabetes mellitus
COPD
Current alcohol abuse
Current smoker
Sepsis-induced hypertension
ARDSa
DIC
Acute oliguria
Lactic acidosis
Pulse, beats/min, mean (SD)
Respiratory rate, breaths/min, mean (SD)
APACHE II score, mean (SD)
White blood cell count (⫻109/L), median
247
106
(70)
(30)
240
108
161 (46)
71 (20)
130 (37)
99 (28)
146 (41)
239 (68)
83 (24)
20 (6)
117 (33)
87 (25)
131.9 (22.5)
33.7 (12.1)
24.2 (6.9)
12.4
(69)
(31)
157 (45)
74 (21)
109 (31)
101 (29)
150 (43)
227 (65)
82 (24)
19 (5)
107 (31)
104 (30)
133.6 (22.7)
33.3 (11.2)
24.3 (7.5)
12.3
COPD, chronic obstructive pulmonary disease; ARDS, acute respiratory distress syndrome; DIC,
disseminated intravascular coagulation; APACHE, Acute Physiology and Chronic Health Evaluation.
a
ARDS is defined as (in the absence of overt cardiac disease): PAO2/FIO2 ratio ⱕ200; diffuse bilateral
infiltrates on chest radiograph; requirement for mechanical ventilation with positive end-expiratory
pressure ⱖ5; pulmonary artery occlusion pressure ⬍19 mm Hg if a pulmonary artery catheter was in
place. The diagnosis of protocol-defined ARDS was made on the day all four criteria were met.
Table 2. Most frequently reported respiratory pathogens in baseline blood cultures
Streptococcus pneumoniae
Staphylococcus aureus
Escherichia coli
Pseudomonas aeruginosa
Staphylococcus species
Streptococcus group A
Placebo
(n ⫽ 353)
Filgrastim
(n ⫽ 348)
70 (20.5)
21 (6.1)
13 (3.8)
9 (2.6)
8 (2.3)
3 (0.8)
60 (18.1)
16 (4.8)
5 (1.5)
2 (0.6)
2 (0.6)
4 (1.2)
Given as number (percent).
Table 3. Most frequently reported respiratory pathogens in baseline sputum cultures
Streptococcus pneumoniae
Staphylococcus aureus
Haemophilus influenzae
Pseudomonas aeruginosa
Escherichia coli
Klebsiella pneumoniae
Streptococcus beta hemolytic
Neisseria species
Moraxella catarrhalis
Enterobacter cloacae
Serratia marcescens
Enterobacter aerogenes
Streptococcus group B
Staphylococcus species
Proteus mirabilis
Given as number (percent).
Crit Care Med 2003 Vol. 31, No. 2
Placebo
(n ⫽ 353)
Filgrastim
(n ⫽ 348)
76 (23.8)
74 (23.2)
39 (12.2)
24 (7.5)
21 (6.6)
11 (3.4)
10 (3.1)
8 (2.5)
7 (2.2)
6 (1.8)
6 (1.8)
6 (1.8)
5 (1.5)
4 (1.2)
4 (1.2)
70 (22.8)
55 (17.9)
29 (9.4)
23 (7.4)
15 (4.8)
15 (4.8)
6 (1.9)
2 (0.6)
6 (1.9)
6 (1.9)
5 (1.6)
3 (0.9)
4 (1.3)
8 (2.6)
3 (0.9)
Efficacy End Points. A total of 191
patients died during the study, 90 in the
placebo group and 101 in the filgrastim
group. The occurrence of mortality
through day 29 was 25.5% in the placebo
group and 29.0% in the filgrastim group.
This difference between the two groups
was not statistically significant (p ⫽ .383,
Cochran-Mantel-Haenszel test stratified
by pooled center). In patients with community-acquired pneumonia, the mortality rates were 24% (placebo) and 28%
(filgrastim) compared with patients with
nosocomial pneumonia with 31% (placebo) and 33% (filgrastim), respectively.
These rates were not significantly different for the filgrastim and placebo groups
in either form of severe pneumonia.
The median time to death was ⬎28
days for both the placebo and filgrastim
groups (Fig. 2). The distributions were
not significantly different between the
treatment groups (p ⫽ .261, log-rank
test). Each covariate effect was significant: age (p ⫽ .0001), number of endorgan dysfunctions (p ⫽ .0001), bacteremia (p ⫽ .0157), and Acute Physiology
and Chronic Health Evaluation II Score
(p ⫽ .0004).
The overall occurrence of new endorgan dysfunctions was 40.0% in the placebo group and 42.6% in the filgrastim
group (p ⫽ .524 according to CochranMantel-Haenszel test stratified by pooled
center). Among antibiotics administered
during the study, only the use of quinolones in the filgrastim-treated group
showed a strong trend toward reduced
28-day mortality. With any use of quinolones from day 1 to day 5, there was 40%
mortality in quinolone-placebo group
(31/78) compared with 29% mortality in
the quinolone-filgrastim group (18/62; p
⫽ .187; 95% confidence interval, 0.45–
1.18). This represented an absolute decrease in mortality of 11% and a relative
risk of .73 (a 27% rate reduction). No
other trends were noted for other antibiotics between the placebo and filgrastimtreated groups.
A total of 686 patients (98%) were in
an intensive care unit at study entry (placebo: n ⫽ 345, 97.7%; filgrastim: n ⫽
341, 97.9%). The median time to intensive care unit discharge was 12 days in
both groups (Fig. 3). The distributions
369
Figure 1. White blood cell response as median white blood cell count by study day. Median peak white
blood cell count was 19.3 ⫻ 109 cells/L for placebo and 31.7 ⫻ 109 cells/L for filgrastim. Median study
day of peak white blood cell count was day 9 for placebo (open circles) and day 4 for filgrastim (filled
circles). The number of patients per treatment group per day is given at bottom of figure.
were not statistically significantly different between the treatment groups (p ⫽
.780, log-rank test). No statistically significant difference was noted in days on
mechanical ventilatory support (9.3 days
in the placebo group and 9.6 days in the
filgrastim-treated group; p ⫽ .835, logrank test).
Safety End Points. A total of 2,517
adverse events were reported among 336
(95%) evaluable patients receiving placebo, and 2,484 adverse events were reported among 338 (97%) evaluable patients receiving filgrastim. Adverse events
were generally similar in type and frequency between the two treatment
groups. Severe adverse events were comparable between the two treatment
groups: 128 (36%) patients in the placebo
group and 137 (39%) patients in the filgrastim group had severe or life-threatening adverse events. The most frequent
severe or life-threatening adverse events
reported with at least a 2% difference in
frequency between the two treatment
groups were renal insufficiency, acute respiratory distress syndrome, hypotension,
cardiac arrest, and acidosis.
No clinically significant effect of filgrastim on laboratory values (other
than white blood cell count) was observed. Filgrastim appeared to be safe
and well tolerated when administered
to 347 patients with pneumonia and
severe sepsis.
DISCUSSION
Figure 2. Kaplan-Meier plot and log-rank test for time to death (expressed as percentage of patients
surviving). Log-rank test, 1.262; p ⫽ .261. Open circles, placebo (n ⫽ 353); filled circles, filgrastim (n
⫽ 348).
Figure 3. Kaplan-Meier plot and log-rank test for time to intensive care unit discharge (expressed as
percentage of patients remaining in the intensive care unit). Log-rank test, 0.078; p ⫽ .789. Open
circles, placebo (n ⫽ 345); filled circles, filgrastim (n ⫽ 341).
370
Nosocomial pneumonia is frequently a
fatal complication in hospitalized patients, particularly those who require mechanical ventilatory support for their
medical problems (8, 9), and communityacquired pneumonia is a major health
problem in Western Europe and North
America. In the United States, pneumonia is the leading cause of death from an
infectious disease and the sixth highest
cause of mortality from any cause (10).
Mortality rates are particularly high in
elderly patients with community-acquired pneumonia of any etiology, even if
not complicated by bacteremia (11).
Younger patients with pneumococcal
pneumonia and sepsis also suffer high
mortalities, usually in excess of 20% (12).
Improved therapeutic strategies are
needed for patients who are severely ill
with pneumonia and sepsis. A study that
is focused on bacterial pneumonia with
strict case definitions and entry criteria
has the virtue of less heterogeneity of the
Crit Care Med 2003 Vol. 31, No. 2
study patients and fewer microbial etiologies to be covered by antimicrobial
treatment. This contrasts with the many
published studies of severe sepsis in
which at least 10% of the patients had no
identifiable source or etiology for sepsis
and may not have had infection as the
cause of the systemic inflammatory response (13).
This study was conducted to determine whether filgrastim, when added to a
course of intravenous antibiotics and
other standard management strategies,
would reduce the rate of mortality in
hospitalized patients with severe pneumonia and sepsis. Filgrastim administration increased white blood cell counts to
a median peak of 31.7 ⫻ 109 cells/L from
a baseline of 12.3 ⫻ 109 cells/L, and were
significantly higher than in the placebo
group. Although the study drug was administered in a blinded fashion, this increase in white blood cell count could be
considered to limit the blinding of the
study. Because the outcome was not different between the two groups, this
study-design limitation is negligible. The
median age, nature and prevalence of comorbid conditions, severity of illness,
rates of bacteremia, prevalence of organ
failures, and the need for supportive mechanical ventilatory support at study entry were well balanced and similar to data
from other published large studies (8,
14). Mortality, the primary efficacy end
point, was not significantly different between the placebo and filgrastim groups,
regardless of whether the cause was community-acquired pneumonia or nosocomial pneumonia. In addition, no significant between-group differences were
found for the secondary efficacy end
points, including the occurrence of endorgan dysfunction, days on mechanical
ventilatory support, time to death, and
time to intensive care unit discharge. Severe adverse events were comparable between the two study groups. Acute respiratory distress syndrome was not
unexpected in this patient population,
and its frequency was within historical
norms in both groups. Acute respiratory
distress syndrome was reported in 41
(12%) patients in the placebo group and
56 (16%) patients in the filgrastimtreated group.
A number of reasons are possible for
this failure to detect any clinical benefit
from adjunctive filgrastim treatment of a
severe infection in which phagocytes are
paramount in host defense. These might
include faulty validity of the underlying
Crit Care Med 2003 Vol. 31, No. 2
hypothesis, inadequate dosage or biological activity of exogenously administered
G-CSF, improper study design, or failed
execution.
It is known that the amount of circulating endogenous G-CSF increases rapidly in patients with bacterial infections,
and most evidence supports a primary
role for this growth factor in generating
the neutrophilic leukocytosis in such patients (15). Particularly high levels of endogenous G-CSF have been reported in
patients with sepsis and septic shock;
these paralleled similar elevations in interleukin-6 and leukemia inhibitor factor
(16). The prompt and significant elevation of absolute neutrophil count in the
filgrastim recipients in this present study
indicates that the dosage chosen had appreciable biological activity, although a
lack of augmentation of other known
phagocytic activities in the setting of severe sepsis cannot be excluded (17).
The choice of the antibiotic therapy
that is paired with filgrastim administration may be important for optimizing favorable anti-infective interactions in
these severely ill patients with sepsis arising from the lungs. As a class, quinolones
exhibit excellent tissue and cell penetration compared with ␤-lactams (18). In
addition, filgrastim has been reported to
increase the intracellular uptake by neutrophils of some antibiotics, including
ciprofloxacin (19). Our results indicate
that patients receiving filgrastim and antibiotic therapy that included a quinolone
for their initial treatment showed a
strong trend toward decreased mortality.
It is conceivable that if quinolones had
been used more broadly, the increased
numbers of patients treated may have
been sufficient for a statistically significant result.
Regarding the study design, the timing of the administration of filgrastim
may have been too late in the course to
alter the clinical outcome of these severely ill patients. In nearly all the preclinical studies of challenge models with
bacteria or bacterial endotoxin, filgrastim
was most beneficial when given either
simultaneously with or before the infectious agent or endotoxin. Much less benefit was observed when filgrastim administration was delayed until well after
initiation of the experimental infection
(5). In the published clinical studies of
filgrastim as an adjunct in the management of bacterial infections, beneficial results were most consistently observed
when filgrastim was given as prophylaxis
D
elaying the administration of
filgrastim to pa-
tients with severe infection
until a time that is well after
initiation of antimicrobial
treatment and the development of organ failures and
systemic immunosuppression may provide little additional anti-infective benefit
and is unlikely to restore
systemic immunity.
to patients at high risk to develop postoperative pneumonia or soft-tissue infections (20, 21) or when administered early
in the management of nonseptic diabetic
patients with localized soft-tissue infection (22). In a prospective, placebocontrolled trial of filgrastim in patients
with bacterial pneumonia, only patients
with multilobar disease without severe
sepsis or multiple organ failures exhibited a strong trend toward reduced mortality (6).
After this study was designed, additional knowledge emerged about the nature of the dynamic balance between systemic inflammatory and counterinflammatory mediators that evolves
during the course of severe sepsis. Although the initial systemic responses to
severe infection are accompanied by
sharp increases in circulating inflammatory mediators, later in the course, the
balance characteristically shifts toward
the anti-inflammatory and immunosuppressive spectrum (23–25). In fact, the
degree of increase in plasma concentrations of a variety of endogenous antiinflammatory mediators correlates
strongly with adverse outcomes (26, 27).
Furthermore, the extent of systemic immunosuppression, as measured by the reduced expression of human leukocyte antigen-DR on circulating mononuclear
cells, also correlates directly with poor
outcomes (28).
371
Several cytokines and growth factors,
including G-CSF, appear to have a key
role in the host’s attempts to restore systemic homeostasis in severe sepsis. In so
doing, their systemic functions may be
quite different from their actions at local
sites of infection (29 –31). The initial responses at the local infection site are predominantly inflammatory and devoted to
eliminating the offending pathogens. In
contrast, the early systemic responses are
usually anti-inflammatory and focus on
restricting some of the damaging systemic consequences of severe local infection (32). With respect to G-CSF, its enhanced production in response to an
infectious challenge is critical for recruiting and optimizing phagocytic activity at
local sites, which are involved in the removal of offending pathogens (15, 17,
32). Conversely, the major systemic effects of G-CSF are both anti-inflammatory and immunosuppressive and appear
to be focused on inhibition of the production and actions of inflammatory cytokines, as well as in expanding a T-helper
lymphocyte response that might eventually lead to the production of specific
antibodies to neutralize microbial pathogenic factors (31, 33–35).
2.
3.
4.
5.
6.
7.
8.
9.
CONCLUSIONS
Delaying the administration of filgrastim to patients with severe infection until
a time that is well after initiation of antimicrobial treatment and the development of organ failures and systemic immunosuppression may provide little
additional anti-infective benefit and is
unlikely to restore systemic immunity.
Further studies using these principles to
explore the timing and ability of filgrastim to ameliorate or prevent the consequences of severe bacterial infection will
be of great interest.
10.
11.
12.
13.
ACKNOWLEDGMENTS
We thank David Dale, MD, and
Thomas Martin, MD, for helpful discussions and suggestions. Mitzi Armstrong,
Naree Sukumoijia,; and the G-ID Clinical
Study Management Team at Amgen were
instrumental in the completion of this
study. MaryAnn Foote, PhD, assisted with
the writing of this manuscript.
14.
15.
16.
REFERENCES
1. Angus DC, Linde-Zwirble WT, Lidicker J, et
al: Epidemiology of severe sepsis in the
United States: Analysis of incidence, out-
372
come, and associated costs of care. Crit Care
Med 2001; 29:1303–1310
Zeni F, Freeman B, Natanson C: Antiinflammatory therapies to treat sepsis and
septic shock: A reassessment. Crit Care Med
1997; 25:1095–1100
Wheeler AP, Bernard GR: Treating patients
with severe sepsis. N Engl J Med 1999; 340:
207–214
Marston BJ, Plouffe JF, File TM Jr, et al:
Incidence of community-acquired pneumonia requiring hospitalization: Results of a
population-based active surveillance study in
Ohio. The Community-Based Pneumonia Incidence Study Group. Arch Intern Med 1997;
157:1709 –1718
Nelson S: A question of balance. Am J Respir
Crit Care Med 1999; 159:1365–1367
Nelson S, Belknap SM, Carlson RW, et al: A
randomized controlled trial of Filgrastim as
an adjunct to antibiotics for treatment of
hospitalized patients with communityacquired pneumonia. J Infect Dis 1998;
8:1075–1080
Wunderink RG, Leeper KV, Schein RM, et al:
Filgrastim in patients with pneumonia and
severe sepsis or septic shock. Chest 2001;
119:523–529
Torres A, Aznar R, Gatell JM, et al: Incidence,
risk, and prognostic factors of nosocomial
pneumonia in mechanically ventilated patients. Am Rev Resp Dis 1990; 142:523–528
Richards MJ, Edwards JR, Culver DH, et al:
Nosocomial infections in medical-surgical
intensive care units in the United States.
Infect Control Hosp Epidemiol 2000; 21:
510 –515
National Center for Health Statistics: Vital
Statistics of the United States. 119th Edition.
Washington, DC, U.S. Census Bureau Statistical Abstract of the United States, 1999, pp
99
Marrie TJ: Community-acquired pneumonia
in the elderly. Clin Infect Dis 2000; 31:
1066 –1078
Feikin DR, Schuchat A, Kolczak M, et al:
Mortality from invasive pneumococcal pneumonia in the era of antibiotic resistance
1995–1997. Am J Public Health 2000; 90:
223–239
Cohen J, Guyatt G, Bernard GR, et al: New
strategies for clinical trials in patients with
sepsis and septic shock. Crit Care Med 2001;
29:880 – 886
Ruiz M, Ewig S, Torres M, et al: Severe community-acquired pneumonia: Risk factors
and follow-up epidemiology. Am J Respir Crit
Care Med 1999; 160:923–929
Dale DC, Liles WC, Summer WR, et al: Review: Granulocyte colony-stimulating factor
role and relationships in infectious diseases.
J Infect Dis 1995; 172:1061–1075
Presneill JJ, Waring PM, Layton JE, et al:
Plasma granulocyte colony-stimulating factor and granulocyte-macrophage colonystimulating factor levels in critical illness,
including sepsis and septic shock: Relation to
disease severity, multiple organ dysfunction,
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
and mortality. Crit Care Med 2000; 28:
2344 –2354
Root RK, Dale DC: Granulocyte colonystimulating factor and granulocyte-macrophage colony-stimulating factor: Comparisons and potential for use in the treatment
of infections in nonneutropenic patients.
J Infect Dis 1999; 179:S342–S352
Moellering RC Jr: The place of quinolones in
everyday clinical practice. Chemotherapy
1996; 42:54 – 61
McKenna PH, Nelson S, Andreson J: Filgrastim enhances ciprofloxacin uptake and bactericidal activity of human neutrophils in
vitro. Abstr. Am J Crit Care Med 1996; 153S:
A535
Schafer H, Hubel K, Bohlen H, et al: Perioperative treatment with filgrastim stimulates
granulocyte function and reduces infectious
complications after esophagectomy. Ann
Hematol 2000; 79:143–151
Wenisch C, Werkgartner T, Sailer H, et al:
Effect of preoperative prophylaxis with filgrastim in cancer neck dissection. Eur J Clin
Invest 2000; 30:460 – 466
Gough A, Clapperton M, Rolando N, et al:
Randomised placebo-controlled trial of granulocyte-colony stimulating factor in diabetic
foot infection. Lancet 1997; 350:855– 859
Bone RC, Grodzin CJ, Balk RA: Sepsis: A new
hypothesis for pathogenesis of the disease
process. Chest 1997; 112:235–243
Dinarello CA: Proinflammatory and antiinflammatory cytokines as mediators in the
pathogenesis of septic shock. Chest 1997;
112:321S–329S
Rangel-Frausto SM, Pittet D, Hwang T, et al:
The dynamics of disease progression in sepsis: Markov modeling describing the natural
history and likely impact of effective antisepsis agents. Clin Infect Dis 1998; 27:185–190
Wray GM, Hinds CJ: Determinants of outcome from sepsis and septic shock. In: Yearbook of Intensive Care and Emergency Medicine. Vincent JL (Ed). Berlin, SpringerVerlag, 1997, pp 168 –179
Gogos C, Drosou E, Bassarus HP, et al: Proversus anti-inflammatory cytokine profile in
patients with severe sepsis: A marker for
prognosis and future therapeutic options.
J Infect Dis 2000; 181:176 –180
Docke WD, Randow F, Syrbe U, et al: Monocyte deactivation in septic patients: Restoration by IFN-␥ treatment. Nat Med 1997;
3:678 – 681
van der Poll T, van Deventer SJ: Cytokines
and anticytokines in the pathogenesis of sepsis. Infect Dis Clin North Am 1999; 13:
413– 426
Munford RS, Pugin J: The crucial role of
systemic responses in the innate (nonadaptive) host defense. J Endotoxin Res
2001; 7:327–332
Hartung T, Docke WD, Handschuh DS, et al:
Effect of filgrastim treatment on inflammatory cytokines and lymphocyte functions.
Clin Pharmacol Ther 1999; 66:415– 424,
Zhang P, Summer WR, Bagby GJ, et al: In-
Crit Care Med 2003 Vol. 31, No. 2
nate immunity and pulmonary host defense.
Immunol Rev 2000; 173:39 –51
33. Boneberg EM, Hareng L, Gantner F, et al:
Human monocytes express functional receptors for granulocyte colony-stimulating factor that mediate suppression of monokines
and interferon-␥. Blood 2000; 95:270 –276
34. Sloand EM, Kim S, Maciejewski JP, et al:
Pharmacologic doses of granulocyte colonystimulating factor affect cytokine production
by lymphocytes in vitro and in vivo. Blood
2000; 95:2269 –2274
35. Arpinati M, Green CL, Heimfeld S, et al:
Granulocyte colony-stimulating factor mobilizes T helper 2-inducing dendritic cells.
Blood 2000; 95:2484 –2490
APPENDIX
Additional Members of the
Pneumonia Sepsis Study Group
Antonio Anzueto, Audie Murphy Memorial VA Hospital, San Antonio, TX;
Mark Astiz, St. Vincent’s Hospital, New
York, NY; Robert Balk, Rush Presbyterian
Hospital, Chicago, IL; Robert Baughman,
University of Cincinnati Hospital, Cincinnati, OH; Steven Belknap, St. Francis
Medical Center, Peoria, IL; David Bihari,
The St. George Hospital, Kogarah, Australia; Bonnie V. Bock, Private practice,
Newport Beach, CA; Raymond Bracis,
Halliday Medical Center, Portland, OR;
Jonathan Burdon, St. Vincent’s Hospital,
Fitzroy, Australia; Richard Carlson, Maricopa Medical Center, Phoenix, AZ; Jean
Chastre, Hopital Bichat, Paris, France;
Nathan Clumeck, Hopital St. Pierre,
Brussels, Belgium; Lynell Collins, VA
Medical Center, Louisville, KY; Richard
Corbin, Carolinas Medical Center, Charlotte, NC; Gerald Criner, Temple Univer-
Crit Care Med 2003 Vol. 31, No. 2
sity Hospital, Philadelphia, PA; Anthony
Dal Nogare, University of Texas Southwest Medical Center, Dallas, TX; Jane
Dematte-D’Amico, Columbia Michael
Reese Hospital, Chicago, IL; Geoffrey
Dobb, Royal Perth Hospital, Perth, Australia; David Dworzack, St. Joseph’s Hospital, Omaha, NE; Jean-Yves Fagon,
Hopital Broussais, Paris, France; Harrison Farber, Boston Medical Center, Boston, MA; Stanley Fiel, MCP Hahnemann
University, Philadelphia, PA; Simon
Finfer, Royal North Shore Hospital, St.
Leonards, Australia; Eugene Fletcher,
University of Louisville Hospital, Louisville, KY; Marilyn Foreman, Moorehous
School of Medicine, Atlanta, GA; Philip
Fracica, St. Joseph’s Hospital, Phoenix,
AZ; Brad Freeman, Washington University, St. Louis, MO; Barry Fuchs, Allegheny University Hospital, Philadelphia, PA;
David Gelmont, LAC/USC Medical Center,
Los Angeles, CA; Jonathan Gottlieb, Jefferson Medical College, Philadelphia, PA;
Benoit Guery, Hopital de Tourcoing,
Tourcoing, France; Jeffrey Hasday, University of Maryland School of Medicine,
Baltimore, MD; Daren Heyland, Kingston
General Hospital, Kingston, Canada; Hoi
Ho, Texas Tech University Hospital, El
Paso, TX; Andrew Holt, Flinders Medical
Centre, Bedford Park, Australia; Donald
Howard, Wilford Hall Medical Center,
Lackland AFB, TX; William E. Hurford,
Massachusetts General Hospital, Boston,
MA; Luc Huyghens, A Z V U B, Brussels,
Belgium; Robert C. Hyzy, Henry Ford
Hospital, Detroit, MI; David Ingbar, University of Minnesota, Minneapolis, MN;
Monroe Karetzky, Newark Beth Israel,
Newark, NJ; Kim M. Kerr, UCSD Medical
Center, San Diego, CA; Gary Kinasewitz,
VA Medical Center, Oklahoma City, OK;
James A. Kruse, Detroit Receiving Hospital, Detroit, MI; Stephen Lapinsky, Mt.
Sinai Hospital, Toronto, Canada; Michael
Littner, VA Medical Center, Sepulvada,
CA; C. Kees Mahutte, VA Medical Center,
Long Beach, CA; John Marshall, Toronto
General Hospital, Toronto, Canada;
George Matuschak, St. Louis University,
St. Louis, MO; C. David Mazer, St. Michael’s Hospital, Toronto, Canada; Alan
Multz, Long Island Jewish Hospital, New
Hyde Park, NY; Erling Myhre, Lund University Hospital, Lund, Sweden; Karim
Nazer, Danbury Hospital, Danbury, CT;
Mary Therese, O’Donnell INOVA Health
Systems, Falls Church, VA; Dan E. Olson,
Medical College of Ohio, Toledo OH; Lucy
Palmer, SUNY at Stony Brook, Stony
Brook, NY; Martin Phillips, Sir Charles
Gairdner Hospital, Nedlands, Australia;
Susan Pingleton, University of Kansas
Medical Center, Kansas City, KS; Robert
Reynolds, PW Clinical Research,
Asheville, NC; Mark Rumbak, Tampa
General Hospital, Tampa, FL; Raffaele
Schicchitano, Royal Adelaide Hospital,
Adelaide, Australia; Terry Smith, Sunnybrook Health Sciences Center, North
York, Canada; Guy Soo Hoo, VA Medical
Center, Los Angeles, CA; Susan Stein, VA
Medical Center, Sepulvada, CA; James
Tan, Akron City Hospital, Akron, OH;
Brian White, Comprehensive Research
Services, Mogadore, OH; John Wilson,
The Alfred Healthcare Group, Phahran,
Australia; Richard Wunderink, Methodist
Healthcare, Memphis, TN; Iven Young,
Royal Prince Alfred Hospital, Camperdown, Australia; Marcus Zervos, William
Beaumont Hospital, Royal Oak, MI
373