critical care

Sepsis: A New Hypothesis for

Pathogenesis of the Disease Process*
Roger C. Bone, MD, PhD (honorary), Master FCCP;' Charles]. Grodzin, MD;
and Robert A Balk, MD, FCCP

(CHEST 1997; 112:235-43) sepsis and its sequelae, innovative therapies were
developed and clinical trials were begun. Although
Abbreviations: CARS =compensatory anti-inflammatory re- these trials were of the most advanced experimental
sponse syndrome; IFN =interferon; IL=interleukin; MARS= design, double-blind, randomized, and placebo con-
mixed antagonists response syndrome; MODS=multiple organ
dysfunction syndrome; SIRS=systemic inflammatory response
trolled, all such sepsis trials thus far have failed to
syndrome; TNF=tumor necrosis factor show efficacy or have had harmful, ambiguous, or
negative results. 4 Pharmacologic interventions to
date have not improved the outcome in sepsis and

SThe
epsis is the systemic response to severe infection.
incidence of sepsis continues to increase.
SIRS. The trials have shown how effective certain
agents can be at the cellular or animal model stage
Sepsis and its sequelae are the leading causes of but how ineffective these same agents can be when
death in medical and surgical ICUs. 1 •2 According to applied in clinical trials. 4
the Centers for Disease Control and Prevention, the While trials addressed the proinflammatory phase
incidence of sepsis continues to increase and is now of sepsis and SIRS, there was no evidence that the
the third leading cause of infectious death (Fig 1). proinflammatory phase was dominant when drugs
Sepsis and its sequelae represent progressive were given. This may mean more to us as we learn
stages of the same illness-a systemic response to more about compensatory anti-inflammatory re-
infection mediated via macrophage-derived cyto- sponses and mixed proinflammatory and anti-inflam-
kines that target end-organ receptors in response to matory responses in the human with sepsis. The
injury or infection. Much confusion has existed failed initial clinical trials tested efficacy of clinical
regarding terminology for sepsis. An American Col- trials for sepsis and provided some insight into the
lege of Chest Physicians/Society of Critical Care complexity of the immuno-inflammatory cascade.
Medicine Consensus Conference 3 held in 1991 This article looks at what we know about this com-
agreed to a new set of definitions that could be plex immuno-inflammatory cascade, and a new hy-
readily applied to patients in different stages of pothesis to relate it to sepsis.
sepsis (Table 1). New discoveries made in the last
several years have validated the conceptual appro-
priateness of these terms, which has led to wide SEPSIS, SIRS, CARS, AND MARS
acceptance. However, new discoveries also suggest
that we need to push these concepts further. When the American College of Chest Physicians
The pathophysiologic state of the systemic inflam- and Society of Critical Care Medicine convened a
matory response syndrome (SIRS) has been studied Consensus Conference in 1991 to address the prob-
extensively. We characterize SIRS as an abnormal lem of confusion over use of proper terms and
generalized inflammatory reaction in organs remote definitions, the terms bacteremia, septicemia, sepsis,
from the initial insult. When the process is due to an sepsis syndrome, and septic shock were being used
infection, the terms sepsis and SIRS are synony- almost interchangeably, which led to confusion and
mous. On the basis of the current understanding of imprecise understanding of sepsis and related disor-
ders. Members of the Consensus Conference agreed
*From the Department of Internal Medicine, Sections of Pulmo- to a new set of definitions that could be readily
nary and Critical Care Medicine, Rush-Presbyterian-St. Luke's applied to patients in different stages of sepsis:
Medical Center, Rush Medical College, Chicago.
1
Deceased. bacteremia, SIRS, sepsis, severe sepsis, septic shock,
Manuscript received April 28, 1997; accepted April 29. and multiple organ dysfunction syndrome (MODS).

CHEST/112/1 /JULY, 1997 235

 30

25
0 • 1980
0
0 20 D 1992

--
0
0
,... 15
( /)
.r::: 10
m
(1)
0 5

0
Lung AIDS Sepsis Urinary Heart Hepato TB
tract biliary
FIGURE l. Leading causes o f infectious death according to ll, c Centers for Disease Control and
Prevention .

We propose now to add the compensatory anti- in those disorders that are often associated with
inflammatmy response syndrome (CARS ), and organ dysfunction, the pattern of systemic cytokine
mixed antagonists response syndrome (MARS ) to release is dissimilar. How, then, can SIRS and
this set of clinical deflnitions (Table 2). MODS be explained ?
Multiple organ d ysfunction occurs in about 30% of
patients with sepsis, and it also can be found in
trauma patients, patients with acute pancreatitis and
THE CYTOKINE CASCADE
othe r diseases such as systemic vasculitides, and in
burn victims. 5 · 9 How dysfunction of multiple organs The systemic response to infection is mediated via
can be produced by such disparate disorders puzzled the macrophage-derived cytokines that target end-
clinicians and investigators for years. Almost a d e- organ receptors in response to injury or infection.
cade ago, it was suggested that multiple organ The inflammatory response to infection or injury is a
dysfunction may result not from infection per se, but highly conserved and regulated reaction of the or-
from a generalized inflammatory reaction.l 0 Evi- ganism . After recognition that a response is required,
dence today suggests that a massive inflammatory the organism (eg, a human being) produces soluble
reaction resulting from systemic cytokine release is protein and lipid proinflammatmy molecules that
the common pathway underlying multiple organ activate cellular defenses, then produces similar
dysfunction. Also, it is now known th at most patients anti-inflammatoty molecules to attenuate and halt
have evidence of dysfunction in one or more organs the proinflamm atmy response. Molecules known or
long before organ failure develops. presumed at this time to be proinflammatory and
Unfortunately, th e more we learn about this in- anti-inflammatory are listed in Table 3. Presumption
flamm atory response, th e more difficult it becomes of activity is basedon data of varyi ng qu ality; it is
to pinpoint a speciflc cytokine, or a specific reaction, likely that some molecules will eventu ally drop from
as the "cause" of SIRS . Indeed, it has become clear this list, and others will be added.
that cytokine release i s a normal, healt hy part of the Normally cytokine response is regulated b y the
body's r esponse to insult or infection. Cytokines are intricate nel:\vork of proinflamm atory and anti-in-
highly pleiotropic, and they appear capable of pro- flamm atmy mediators. The initial inflammatory re-
ducing markedly different effects depending on the sponse is kept in check b y down-regulating produc-
nearby hormonal milieu. Fmthennore, the body has tion and counteracting the effects of cytokines
a highly complex, rigidly regulated n etwork of recep- already produced. The picture that emerges from
tor antagonists and other r egulatory agents that analysis of data from patients with sepsis is that a
continuously modulate the effects of cytokine re- complex mixture of proinflammatory and anti-in-
lease . Adding to our confusion is the fact that flammatory molecules may be present.11 · 12 Standard
systemic cytokin e release can occur in a variety of pathophysiologic models of sepsis do not explain
disorders without leading to organ dysfunction. Even such a picture.l3

236 Critical Care

 Table !-Standard Definitions for Sepsis and Organ Table 2-Proposed Definitions for Sepsis and Organ
Failure Failure

Terminology Definition CARS HLA-DR on monocytes <30% and diminished

ability of monocytes to produce inflam matory
Infection Microbial phenomenon characterized by a n
cytokines, such as TNF-a or IL-6 (Kox WJ, Bone
inflammatory response to the presence of
RC , Krausch D, et a!. Arch Intern Mecl 1997; 157:
microorganisms or the invasion of normally
.389-93).
sterile host tissue by those organisms.
MARS Features of SIRS in a patient with CARS.
Bacteremia Presence of viable bacteria in the blood.
SIRS The systemic inflammatory response to a \vide
variety of severe clinical insults, manifested by
two or more of the following conditions: (1)
temperature >38°C or <36°C; (2) heart rate proinflammatory reaction (SIRS) or a compensatory
>90 beats/min; (3) respiratory rate >20 anti-inflammatory reaction (CARS) will ensue. A
breaths/min or PaC0 2 <32 mm Hg; and (4)
range of clinical sequelae may then follow. These
WBC count > 12,000/mm 3 , <4,000/mm 3 , or
> 10% immature (band ) form s. sequelae can be remembered by using the mne-
Sepsis The systemic inflammatory response to infection. monic CHAOS (Fig 2) .
In association with infection, manifestations of Currently, our concept of the pathogenesis of
sepsis are the same as those previously defined sepsis is undergoing evolution, based in part on
for SIRS. It should be determined whether
animal models or human exposure to endotoxin, in
they are a part of the direct systemic response
to the presence of an infectious process and hopes of finding a "magic bullet" for sepsis. 4 The
represent an acute alteration from baseline in magic bullet to definitively treat sepsis has been
th e absence of other known causes for such vigorously sought, but has not been found. More
abnormalities. than a dozen pharmacologic magic bullet candidates
Severe sepsis Sepsis associated with organ dysfunction,
have failed to improve outcome of sepsis in random-
hypoperfusion, or hypotension. Hypoperfusion
and perfusion abnormalities may include, but ized, placebo-controlled clinical trials . Magic bullet
are not limited to, lactic acidosis, oligUiia, or an trials were based on an assumption that antagonism
acute alteration in mental status. of a single proinflammatory mediator can modulate
Septic shock A subset of severe sepsis and defined as sepsis- the cascade of events that constitutes sepsis in a
induced hypotension despite adequate fluid
heterogeneous group of patients. Hindsight indicates
resuscitation along with the presence of
perfusion abnormaliti es that may include, but that this was simplistic, but evidence available at that
are not limited to, lactic acidosis, oligUiia, or an time was interpreted to suggest that such a strategy
acute alteration in mental status. Patients should be successful. Similarly, available evidence
receiving inotropic or vasopressor agents may suggested that neutralization of endotoxin would
no longer b e hypotensive by the time they
prevent the proinflammatory response, but clinical
manifest hypoperfusion abnormalities or organ
dysfunction, yet they would still be considered trials of monoclonal antiendotoxin antibodies were
to have septic shock. not successful. 15 - 18 The design of clinical trials man-
MODS Presence of altered organ function in an acutely dated that patients enrolled in the trial meet some
ill patient such that homeostasis cannot be criteria for having sepsis, ignoring any preexisting
maintained without intervention.
condition that might have induced proinflammatory
Sepsis-induced A systolic BP <90 mm Hg or a reduction of ~40
hypotension mm Hg from baseline in the absence of other mediators. 19 Animal studies suggested that an endo-
causes for hypotension. toxin-neutralizing drug should be given before or
shortly after the inflammatory stimulus, but interspe-
cies variation in immunomodulation and other dif-
ferences in animal models compared to the critically
These mediators initiate overlapping processes ill ICU patient make it difficult to generalize animal
that directly influence the endothelium, cardiovascu- studies to humans. 2 o
lar, hemodynamic, and coagulation mechanisms. The
release of many of these vasoregulators is often local.
Evolving concepts of the septic response give more WHY PREVIOUS THEORIES WERE INADEQUATE
weight to the importance of local cytokine produc-
tion, not in contradistinction to systemic production We have understood that a massive inflammatory
but as part of the total septic-response picture. 4 · 14 reaction underlies both SIRS and MODS, but now
The duration of illness also may alter the mix of we must understand that this reaction is only half the
mediators, leading to a state of metabolic disorders picture. It is now clear that quite rapidly after the
in which the body has no control over its own first proinflammatory mediators are released, the
inflammatory response. If balance cannot be estab- body mounts a compensatory anti-inflammatory re-
lished and homeostasis is not restored, a massive action (CARS) to the initial proinflammatory re-

CHEST /112/1 I JULY, 1997 237

 Table 3-Partial List of Proinflammatory and Anti-inflammatory Molecules
l'roinflammatory Molecules Anti-inflammat01y Molecules
TN F-a Thromboxane IL-l ra
IL-113 Platelet activatin g factor IL-4
IL-2 Soluble adhesion molecules IL-10
IL-6 Vasoactive neuropeptides IL-13
IL-8 Phospholipase A2 Type II I L-1 receptor
IL-15 Tyrosine kinase Transforming growth factor-13
Neutrophil elastase Plas minogen activator inhibitor- I Epinephrine
IFN--y Free radical generation Soluble TNF-a receptors
Protein kinase Neopterin Leukotriene B4 -receptor antagon ism
MCP-1* CD14 Soluble re combinant CD-14
MCP-2 Prostacyclin LPS binding protein*
Leukemia inhibitory factor Prostaglandins
(D-factor)
*MCP = monocyte chemoattractant protein; LPS = Iipopolysaccharide.

sponseP The anti-inflammatory reaction may be as matory mediators and homeostasis is restored (Fig
large as, and sometimes even larger than, the proin- 2) . In some patients, however, a variety of forces
flammatory response. The goal of this anti-inflam- conspire to upset this balance, resulting in SIRS and
matmy reaction is to down-regulate synthesis of MODS.
proinflammatory mediators and to modulate their The theories put forth to explain the development
effects, thereby restoring homeostasis. of SIRS have generally not taken this compensatory
It has recently become possible to differentiate anti-inflammatOI)' reaction into consideration. Many
ongoing CARS from SIRS immunophysiology. of the anti-inflammatory mediators were discovered
Zedler et al21 detailed a technique of stimulating and characterized only in the last few years, and to
peripheral blood mononuclear cells from severely some extent, this may have led to overstatement of
injured burn patients for the purpose of cell surface
the dangers presented by proinflammatory media-
antigen staining and intracellular interferon-gamma
tors . It might almost be said that proinflammatmy
(IFN-)') and interleukin-4 (IL-4) detection. IL-4, an
mediators became "bad guys ," without taking into
anti-inflammatory cytokine, was found in excess (el-
account that excessive levels can be harmful, but
evated 16-fold) in the presence of downregulated
IL-2 and IFN-)'. IL-4 thus served as a marker for the lower levels are required to combat pathogenic
"THrTH 2 switch," a major characteristic of the organisms and to promote healing.
CARS response to injury (THis the T h elper cell) . In Most of the evidence for the role of proinflamma-
most healthy persons, the body is able to achieve a tory mediators in the pathogenesis of SIRS and
balance between proinflammatory and anti-inflam- MODS came from studies using animal models,
experiments in which endotoxin or proinflammatory
mediators were injected into human volunteers, and
analysis of serum levels of proinflammato1y media-
C Cardiovascular compromise (usually manifesting tors in patients \Vith sepsis, burn injury, or other
as shock; in this setting SIRS predominates). severe injuries (Fig 3) . We now know that these
studies may not truly reflect what happens in criti-
H Homeostasis (return to health; this represents a cally ill patients with sepsis or SIRS. As noted earlier,
balance of SIRS and CARS). a marked interspecies variation in cytokine release
makes it difficult to extrapolate results of animal
A Apoptosis (neither SIRS nor CARS predominates). studies to humans. More importantly, these experi-
ments were performed on healthy animals and gen-
0 Organ dysfunction (single or multiple;
erally included a relatively short observation peri-
SIRS predominates). od.22 Studies of human volunteers were performed
in healthy subjects; the amount of stimulus injected
S Suppression of the immune system (anergy and/or was sublethal; and, again, the follow-up period was
increased susceptibility to infection ; CARS brief. 23 In contrast, SIRS and MODS develop over
predominates) . time in severely ill or injured patients who have
multiple preexisting disorders.24
FIGURE 2. Mnemonic of CHAOS. Serum levels of immunomodulating mediators

238 Critical Care

 result of an immediate insult such as trauma and not
a consequence of a preexisting condition such as
Infection pancreatitis.19

'
RELATING CLINICAL RESPONSES TO CYTOKINE
CASCADE

Endotoxin and Some patients with sepsis, extensive burns, mas-

sive traumatic injury, or other severe insults show
other microbial toxins little or no evidence of a systemic inflammatory
reaction or organ dysfunction, although their recov-

'
ery may be protracted because of the severity of their
underlying illness. In three other categories, how-
ever, are patients with sepsis or other severe insult
who develop the following: a mild form of SIRS and
Proinflammatory state some evidence of dysfunction in one or two organs
with cytokine release early in their clinical course that usually resolves
rapidly; a massive systemic inflammatory reaction
and developing rapidly after the initial insult, with death
other proinflammatory often following within a few days from profound
shock; and a less severe initial course, but marked
mediators deterioration several days or more after the original
insult, with outright failure of one or more organs

'
and death in some but not all patients.
Clinical trials have usually excluded patients with
mild symptoms of organ dysfunction or symptoms
that last for <24 or 48 h. While this was perhaps
Sepsis/SIRS believed necessary to the design and conduct of
these trials, the underlying hypothesis may have

'
been faulty. We should have looked better at the
proinflammatory and anti-inflammatory response to
severe insult and asked whether the inflammatory
response was of an appropriate magnitude and if it
Shock and multiorgan was appropriately down-regulated. When down-reg-
dysfunction and ulation is not adequate, is there a progression of
severities? Rangel-Frausto et al 28 published the first
possible large study to confirm that patients progress through
death stages of the septic process, from mild to severe.

FIGURE 3. Old paradigm for sepsis. A NEW THEORY OF SIRS, CARS, MARS, AND
MODS
Immunomodulation is a complex, overlapping net-
present problems of interpretation because immu- work of interactions among agents that work to-
noassays can detect only free, circulating mediators, gether to overcome severe assaults on the body.
not mediators bound to cells or receptors. 12,2.5,26 Paradoxically, they also work to the body's disadvan-
Therefore, the amount of mediator reported may not tage and cause the disruptions we call SIRS and
be the amount present. Bioassays used to measure MODS. We have presented a hypothesis-based ex-
the functional activity of cytokine often lack speci- planation for the apparently paradoxical events ob-
ficity and may over-report the amount of mediator served in the ctitically ill (Fig 4). The five stages in
present. 27 Other points to consider are the following: the development of multiple organ dysfunction are
( l) analysis of serum level is usually performed once as follows: (1) local reaction at the site of injury or
a day or less often, although mediator release is infection; (2) initial systemic response; (3) massive
phasic; and (2) most analyses have assumed that the systemic inflammation; (4) excessive immunosup-
presence of proinflammatory mediators is a direct pression; and (5) immunologic dissonance. 29

CHEST I 112 I 1 I JULY, 1997 239

 In itial insult
(bacterial,
viral, traumatic,
thennal)

A 0 s
Cardiovascular Hom eo- Apoptosis Organ Suppression
compromise stasis (cell death) dysfunction of the
(shock) immune
system

SIRS CARS and Death w~h SIRS CARS

predominates SIRS minimal predominates predominates
balanced inflammation

FIGURE 4. New concepts for the clinical sequelae of sepsis, SIRS , CARS, and MARS . (This figure is
an adaptation of Figure 1by Bone RC. Sir Isaac Newton, sepsis, SIRS, and CAH.S. Crit Care Med 1996;
24:1125-28.)

Stage 1 substances 12 •30 -32 work to diminish monocytic major

histocompatibility complex class II expression, im-
Prior to development of SIRS or MODS is some
pair antigen presenting activity, and reduce the
insult such as a nidus of infection, a traumatic injury
ability of cells to produce inflammatory cytokines.
(including a surgical wound), a burn injury, or
Local levels of both proinflammatory and anti-in-
pancreatitis that prompts release of a variety of
flammatory mediators can be substantially higher
mediators in the microenvironment. The body's ini-
than are later found syste mically33 -38 (Table 3).
tial response is to induce a proinflammatory state in
which mediators have multiple overlapping effects
Stage 2
designed to limit new damage and to ameliorate
whatever damage has already occurred. They destroy If the original insult is sufficiently severe, first
damaged tissue, promote the growth of new tissue, proinflammatory and later anti -inflammatory media-
and combat pathogenic organisms, neoplastic cells, tors will appear in the systemic circulation via a
and foreign antigens 2 0 ·24 (Table 3). variety of mechanisms. The presence of proinflam-
A compensatory anti-inflammatory response soon matOiy mediators in the circulation is part of the
ensures that the effects of these proinflammatory normal response to infection and serves as a warning
mediators do not become destructive. IL-4, IL-10, signal that the microenvironment cannot control the
lL-11, IL-13, soluble tumor necrosis factor (TNF-a) initiating insult. The proinflammatory mediators
receptors, IL-l receptor antagonists, transforming help recruit neutrophils, T cells and B cells, platelets,
growth factor-[3 , and other, as yet undiscovered and coagulation factors to the site of injury or

240 Critical Care

infection. 24 This cascade stimulates a compensatory Stage 5
systemic anti-inflammatmy response, which nor-
The final stage of MODS is what we have elected
mally quicldy down-regulates the initial proinflam-
to call "immunologic dissonance."54 It is an inappro-
matory response. Few, if any, significant clinical
priate, out-of-balance response of the immunomodu-
signs and symptoms are produced. Organs may be
latory system. In some patients, it results from
affected by the inflammatory cascade, but significant persistent, overwhelming inflammation that may
organ dysfunction is rare. persist in patients with SIRS and MODS, with
increased risk of death.u,.ss-.57
Stage 3 In some patients, persistent immune suppression
causes immunologic dissonance. Studies have shown
Loss of regulation of the proinflammatory re- not only that monocyte deactivation persists in many
sponse results in a massive systemic reaction mani- patients, but that such persistent deactivation greatly
fest as the clinical findings of SIRS. Underlying the increases the risk of death.51 In patients with persis-
clinical findings are pathophysiologic changes that tent immune suppression, the cause of organ failure
include the following: (1) progressive endothelial may be inhibition of the· synthesis of the proinflam-
dysfunction, leading to increased microvascular per- matory agents needed to allow the organs to recover.
meability;39-43 (2) platelet sludging that blocks the In patients with immunologic dissonance, it may be
microcirculation, 44 causing maldistribution of blood possible to regain organ function if the body can
flow and possibly ischemia, which in turn may cause recover its balance.
reperfusion injmy5 and induction of heat shock
proteins;46 (3) activation of the coagulation system
and impairment of the protein C-protein S inhibitory CONCLUSIONS
pathway;47 and (4) profound vasodilation, fluid tran-
Balance between proinflammatory and anti-in-
sudation, and maldistiibution of blood flow may
flammatmy forces could conceivably be lost: ( 1)
result in profound shock. 48 ·49 Organ dysfunction and, when infection, burn injury, hemorrhage, etc, is so
ultimately, failure result from these changes unless severe that the insult alone is sufficient to prompt
homeostasis is quickly restored. SIRS and MODS; or (2) when patients are "pre-
primed" to develop SIRS and MODS by preexisting
Stage 4 severe illness. (3) Most of the preexisting conditions
are associated with abnormal cytokine levels. 19
It is possible that a compensatmy anti-inflamma- The hypothesis of prepriming rests on our under-
tory reaction can be inappropriate, with a resulting standing that a patient at risk for SIRS or MODS
immunosuppression. What some investigators have already has a significant clinical history, and is not
called "immune paralysis,"50·51 and "window of im- clinically analogous to a healthy human volunteer.
munodeficiency,"·52 we describe as "compensatory
anti-inflammatory response syndrome" (CARS). 4
CARS is the body's response to inflammation and is REFERENCES
more than just immune-paralysis. CARS may explain 1 Tran DD, Groenveld AB , van der M eulen J, et !.a Age,
such anomalies as the burn patient's increased sus- chronic disease, sepsis, organ syste m failure, and mortality in
ceptibility to infection and even the anergy of the a medical intensive care unit. Crit Care Med 1990; 18:474-79
pancreatitis patient. 29 ·53 Recently, it has been shown 2 Marshall JC, Cook DJ, Christou NC, et !a. Multiple organ
that treatment of patients with sepsis with IFN--y not dys function score: a reliable descriptor of a complex clinical
outcome. Crit Care Med 1995; 23:1638-52
only restores the HLA-DR expression on monocytes 3 American College of Chest Physicians/Society of Critical Care
but reestablishes the ability of monocytes to secrete Medicine Consensus Conference. Definitions for sepsis and
the cytokines IL-6 and TNF-a. 29 One recent study55 multiple organ failure, and guidelines for the use of innova-
looked at a series of patients admitted to an ICU who tive the rapies in sepsis. Crit Care Med 1992; 20:864-74
met the criteria for SIRS. All patients demonstrated 4 Bone RC. Why sepsis trials htil. JAMA 1996; 276:565-66
5 Livingston DH, Appel SH , Wellhausen SR, et al. Depressed
CARS immunophysiology in that monocyte surface inte rferon-gamma production and monocyte HLA-DR ex-
human leukocyte antigen (HLA)-DR expression was pression after severe injury. Arch Surg 1988; 123:1309-13
reduced to < 30%. The study confirmed that admin- 6 Cheadle WG, Hershman MJ, Wellhausen SR, e t al. Role of
istration of IFN--y-1b could effect up-regulation of monocytic HLA-DR expression following trauma in predict-
monocyte surface HLA-DR expression, restoration ing clinical outcome. In : Faist E, Ninnemann J, Green D , eds.
Immune consequences of trauma, shock and sepsis. New
of monocyte function , and secretion of IL-6 and York: Springer V erl ag, 1989; 119-22
TNF -a. The resultant state rebalances CARS-SIRS 7 Gibbons RA, Ma1tinez OM , Lim OC, et a!. Reduction in
immunophysiologic homeostasis. HLA-DR, HLA-DQ, and HLA-DP exp ression by Leu-M3 +

CHEST I 112 I 1 I JULY, 1997 241

 cells from the peripheral blood of patients with thermal 28 Rangel-Frausto MS, Pittet D, Costigan M, et al. The natural
injury. Injury 1989; 19:263-66 history of the systemic inflammatory response syndrome
8 Wakefield CH, Carey PD, Foulds S, eta!. Changes in major (SIRS): a prospective study. JAMA 1995; 272:117-23
histocompatibility complex class II expression in monocytes 29 Bone RC. Immunologic dissonance: a continuing evolution in
and T cells of patients developing infection after surgery. Br J our understanding of the systemic inflammatory response
Surg 1993; 80:205-09 syndrome (SIRS ) and the multiple organ dysfunction syn-
9 Broader W, Williams D, Pretus H, et a!. Beneficial effect of drome (MODS ). Ann Intern Med 1996; 125:680-87
enhanced macrophage function in the trauma patient. Ann 30 Fisher CJ, Dhainaut JF, Opal SM, eta!. Recombinant human
Smg 1990; 211:605-12 interleukin-1 receptor antagonist in the treatment of patients
10 Goris JA, te Boekhorst TPA, Nuytinck JKS , et a!. Multiple- with sepsis syndrome: results from a randomized, double-
organ failure: generalized autodestructive inflammation ? blind, placebo-controlled trial. JAMA 1994; 271:1836-44
Arch Surg 1985; 120:1109-15 31 Abraham E, Wunderink R, Silverman H, et al. Efficacy and
ll Pinsky MR, Vincent J-L, Deviere J, et a!. Serum cytokine safety of monoclonal antibody to human tumor necrosis
levels in human septic shock: relation to multiple-system factors in patients with sepsis syndrome. JAMA 1995; 273:
organ failure and mortality. Chest 1993; 103:565-75 934-37
12 Dinarello CA, Gelfand JA, Wolff SM. Anticytokine strategies 32 Platzer C, Meisel C ,Vogt K , e tal. Up-regulation of monocytic
in the treatment of the systemic inflammatory response IL-10 by tumor necrosis factors and cAMP elevating drugs.
syndrome. JAMA 1993; 269:1829-35 Intern Immunol 1995; 7:17-23
13 Goldie AS, Fearon KCH, Ross JA, e t al. Natural cytokine 33 Fukushima R, Alexander JW, Cianotti L, et al. Isolated
antagonists and endogenous antiendotoxin core antibodies in pulmonary infection acts as a source of systemic tumor
sepsis syndrome. JAMA 1995; 274:172-77 necrosis factor. Crit Care Med 1994; 22:114-20
14 Moore FA, Moore EE, Read RA. Postinjury multiple organ 34 Ford HR, Hoffman RA, Wing EJ, et al. Characterization of
failure : role of extra-thoracic injury and sepsis in adult wound cytokines in th e sponge matrix model. Arch Surg 1989;
respiratory distress syndrome. New Horizons 1993; 1:538-49 124:1422-28
15 Bone RC, Balk RA, Fein AM , eta!. A second large controlled 35 Meduri GU, Kohler G, Headley S, et al . Inflammatory
clinical study of E5, a monoclonal antibody to endotoxin: cytokines in the BAL of patients with ARDS: persistent
results of a prospective, multicenter, randomized clinical trial. elevation over time predicts a poor outcome. Chest 1995;
Crit Care Med 1995; 23:994-1006 108:1303-14
16 Greenman RI, Schein RM, Martin MA, et al. A controlled 36 Sauder DN, Semple J, Truscott D, e t al. Stimulation of
trial of E5 murine monoclonaligM antibody to endotoxin in muscle protein degradation by murine and human epidermal
the treatment of Gram-negative sepsis. JAMA 1991; 266: cytokines: relationship to thermal injmy. J Invest Dermatol
1097-1102 1986; 87:711-14
17 McCloskey RV, Straube RC, Sanders C, e t a!. Treatment of 37 Kupper TS, Deitch EA, Baker CC, et al. The human burn
septic shock with human monoclonal antibody HA-1A: a wound as a primary source of interleukin-1 activity. Surgery
randomized double-blind, placebo-controlled trial. Ann In- 1986; 100:409-14
tern Med 1994; 121:1-5 38 Puren AJ, Feldman C, Savage N, et al. Patterns of cytokine
18 Ziegler EJ, Fisher CJ, Sprung CL, et al. Treatment of expression in community-acquired pneumonia. Chest 1995;
Gram-negative bacteremia and septic shock with HA-lA 107:1342-49
human monoclonal antibody against endotoxin. N Eng] J Med 39 Tracey KJ, Lowry SF, Ceremi A. Cachectin/TNF-alpha in
1991; 324:429-36 septic shock and adult respiratory distress syndrome [edito-
19 Bone RC. Toward a theory regarding the pathogenesis of th e rial]. Am Rev Respir Dis 1988; 138:1377-79
systemic inflammatory response syndrome: what we do and 40 Stephens KE , Ishikaza A, Larrick JW, et al. Tumor necrosis
do not know about cytokine regulation. Crit Care Med 1996; factor causes increased pulmonary permeability and edema:
24:163-72 comparison to septic acute lung injury. Am Rev Respir Dis
20 Zuckerman SH, Bendele AM. Regulation of' serum tumor 1988; 137:1364-70
necrosis factor in glucocorticoid-sensitive and -resistant ro- 41 Ibbotson GC, Wallace JL. Beneficial effects of prostaglandin
dent endotoxin shock models. Infect Immun 1989; E2 in endotoxin shock are unrelated to effects of PAF -acether
57:3009-13 synthesis. Prostaglandins 1989; 37:237-50
21 Zedler S, Bone RC, v. Donnersmarck GH, et a!. T-cell 42 Lewis RA, Austen KF, Soberrnan RJ. Leukotrienes and other
reactivity and its predictive rol e in immunosuppression after products of the 5-lipoxygenase pathway: biochemistry and
burns (submitted) relation to pathobiology in human diseases. N Eng! J Med
22 Balk RA, Jacobs RF, Tryka AF, e t al. Effect of ibuprofen on 1990; 323:645-55
neutrophil function and acute lung injmy in canine endotoxin 43 Petrak RA, Balk RA, Bone RC. Prostaglandins, cyclo-oxygen-
shock. Crit Care Med 1988; 15:1121-24 ase inhibitors, and thromboxane synthesis inhibitors in the
23 Suffredini AF, Fromm RE, Parker MM , e t a!. The cardiovas- pathogenesis of multiple organ failure. Crit Care Clin 1989;
cular response of normal humans in the administration of 35:303-14
endotoxin. N Eng! J Med 1989; 321:280-87 44 Sigurdsson GH, Christenson JT, Bader el-Rakshy M, et al.
24 Bone RC. The pathogenesis of sepsis. Ann Intern Med 1991 ; Intestin al platelet trapping after traumatic and septic shock:
115:457-69 an early sign of sepsis and multiorgan failure in cJitically ill
25 Munoz C, Carlet J, Fitting C, eta!. Dysregulation of in vitro patients? JAMA 1992; 20:458-67
cytokine production by monocytes during sepsis. J Clin Invest 45 Cipolle MD, Pasquale NM, Cerra FB . Secondary organ
1991; 88:1747-54 dysfunction: from clinical perspectives to molecular media-
26 Keogh C, Fong Y, Marano MA , e tal. Identillcation of a novel tors. Crit Care Clin 1993; 9:261-98
tumor necrosis factor alcachectin from the livers of burned 46 Rinaldo JE, Garry M, Streiter R , etal. Effect of endotoxin-
and infected rats. Arch Surg 1990; 125:79-85 induced cell injury on 70-kD heat shock proteins in bovine
27 Remick DG. Applied molecular biology of sepsis. J Crit Care lung endothelial cells. Am J Respir Cell Mol Bioi 1990;
1995; 10:198-212 3:207-16

242 Critical Care

47 Levi M, ten Cate H, van der Poll T, et a!. Pathogenesis of 53 Hamilton G, Hofbauer S, Hamilton B. TNF-alpha, interleu-
disseminated intravascular coagulation in sepsis. JAMA 1993; kin-6 and parameters of the cellular immune system in
270:975-79 patients with intra-abdominal sepsis. Scand J Infect Dis 1992;
48 Gomez-Jimenez J, Salgado A, Mourelle M, eta!. L-arginine: 24:361-68
nitric oxide pathway in endotoxemia and human septic shock. 54 Kox WJ, Bone RC, Krausch D, et a!. Interferon--y in the
treatment of the compensatory anti-inflammatory response
Crit Care Med 1995; 23:253-58
syndrome (CARS)-a new approach: proof of principle. Arch
49 Miyauchi T, Tomobe Y, Shiba R, et a!. Involvement of Intern Med 1997; 157:389-93
endothelin in the regulation of human vascular tonus. Circu- 55 Fisher CJ Jr, Opal SM, Dhainaut JF, eta!. Influence of an
lation 1990; 81:1874-80 anti-tumor necrosis factor monoclonal antibody on cyto-
50 Randow F, Syrbe U, Meisel C, eta!. Mechanism of endotoxin kine levels in patients with sepsis. Crit Care Med 1993;
desensitization: involvement of interleukin-10 and transform- 21:318-27
ing growth factor beta. J Exp Med 1995; 181:1887-92 56 Dofferhoff AS, Yom VJ, de Vries-Hospers HG, eta!. Patterns
51 Syrbe U, Meinecke A, Platzer C, et a!. Improvement of of cytokines, plasma endotoxin, plasminogen activator inhib-
monocyte function-a new therapeutic approach? In: Sepsis: itor, and acute-phase proteins during the treatment of severe
current perspectives in pathophysiology and therapy. New sepsis in humans. Crit Care Med 1992; 20:185-92
York: Springer Verlag, 1994; 473-500 57 Lesser HG, Gross V, Scheibenbogen C, et a!. Elevation of
52 Mills CD, Caldwell NM, Gann DJ. Evidence of a plasma- serum interleukin-6 concentration precedes acute-phase re-
mediated "window" of immunodeficiency in rats following sponse and reflects severity in acute pancreatitis. Gastroen-
trauma. J Clin Immunol1989; 9:139-50 terology 1991; 101:782-85

CHEST /112/1 I JULY, 1997 243