editorials

VOLUME 101 I NUMBER 6 I JUNE, 1992

knowledge of these events is still incomplete, we have
The ACCP-SCCM Consensus learned enough to make the need for more exacting
Conference on sepsis and Organ definitions apparent.
Failure 7. There have not been enough epidemiologic stud-
ies to evaluate the systemic response to infection and

U nited States Supreme Court Justice Potter Stewart

once wrote, ccI caDt define obscenity, but I know
its consequent sequelae of sepsis and multiple organ
failure.
it when I see it."l Until quite recentl~ that statement Last year, one of us (R. C. B.) published two
could equally well have been applied to sepsis and articles6 ,7 in an initial attempt to establish more
related disorders. Even in four recent multicenter uniform terminology. In subsequent commentaries,
trials, markedly different definitions of sepsis, shock, two of us (C. L. S. and W. J. S.) took issue with many
and organ failure were employed.2-5 ofthe original recommendations. 8 ,9 However, it quickly
From a clinical standpoint, the absence of firm became clear that there was much we agreed on, and
definitions for these disorders has had-until now- that there were many questions we all felt needed to
little practical consequence because the only treat- be raised and answered. For example, is sepsis an
ments available have been antibiotics and supportive appropriate term for a process that may take place in
care. Given the abundance ofnew agents under active the absence of infection? Can shock be defined solely
investigation, however, more precise diagnostic guide- in terms of blood pressure? If the extent of organ
lines are needed to allow us to evaluate effic~ to dysfunction can vary, is it accurate to speak of organ
determine which patients might benefit from such ccfailure"? Each of us has attempted to use current
treatments, and to compare the results of different terminology to address these questions, but we were,
trials. in essence, trying to fit square pegs into round holes.
There are a number of historical reasons why firm Because of this need, a consensus conference, spon-
definitions have been lacking: sored by the American College of Chest Physicians
1. Until roughly 30 years ago, sepsis, septic shock, and the Society of Critical Care Medicine, was held.
and multiple organ failure were rarely seen. Simply On page 1644 of this issue, the consensus statement
put, we could not keep severely ill or injured patients from that conference appears. The participants agreed
alive long enough for these disorders to develop. that two new terms are necessary:
2. Many of the early studies of sepsis focused on 1. Systemic inflammatory response syndrome
patients with Gram-negative bacteremia. However, (SIRS): This term is preferred to sepsis when describ-
we now know that the majority of patients with sepsis ing the widespread inflammation (or clincal response
are not bacteremic. to that inHammation) that can occur in patients with
3. Many of the early studies were conducted in such diverse disorders as infection, pancreatitis, ische-
surgical patients or trauma victims; it was not clear mia, multiple trauma, hemorrhagic shock, and im-
whether the physiologic derangements that occurred munologically mediated organ inju~ The term sepsis,
in these patients were the same as those that occurred a subcategory of the dysfunction newly defined as
in patients with Gram-negative bacteremia. SIRS, should be used only for those patients with
4. The lack ofprecise criteria for the terms infection, documented infection.
sepsis, sepsis syndrome, and septic shock made it 2. Multiple organ dysfunction syndrome (MODS):
difficult to assess the severity of the infectious process The extent of organ dysfunction in patients with SIRS
and the differences between study populations. can vary tremendousl~ both from patient to patient
5. The lack of precise criteria for the term multiple and within the same patient over time. The term
organfailure made it difficult to establish which organs MODS was coined to indicate the wide range of
were affected in patients with sepsis. It also made it severity and the dynamic nature ofthis disorder. There
more difficult to determine whether organ failure was are two relatively distinct (although not mutually
a cause-or a consequence-of sepsis. exclusive) pathways by which MODS can develop: In
6. Only recently has the knowledge of the molec- primary MODS, there is a direct insult to the organ
ular and cellular events that occur in sepsis and its that becomes dysfunctional. Examples of such direct
sequelae begun to shed light on the complex cascade insults include gastric aspiration in the lungs or
of events underlying these disorders. Although our rhabdomyolysis in the kidney. This direct insult causes

CHEST I 101 I 6 I JUNE, 1992 1481

an inHammatory response that is localized, at least in conjunction with the diagnostic criteria established
the beginning, to the affected organ. Secondary for SIRS" MODS, and related disorders-may be the
MODS is a consequence of trauma or infection in one best available method for predicting which patients
part of the system that results in the systemic inHam- are at greatest risk. Obviousl~ research is needed to
matory response and dysfunction oforgans elsewhere. confirm (or refute) this belief and to elucidate how
An understanding of the etiology of organ dysfunction these scoring systems, which were designed to assess
is important in the treatment ofthe underlying disease prognosis in large groups, can be used to predict
often present in MODS. outcome in individual patients.
In addition, the consensus conference agreed on The final task of the consensus conference was to
definitions for the terms bactenmda, sepsis, severe provide recommendations on how these new defini-
sepsis, septic shock, and other related disorders. In tions could be incorporated into the design of clinical
developing these de6nitions, another problem became trials for new treatments that can fight SIRS and
apparent, particularly in light of recent clinical trials: MODS. In response to many ofthe concerns discussed
We currently do not have a good way to predict which above, the conference also sought to provide guide-
patients will develop SIRS or MODS. For example, lines for the design of these trials.
many patients with severe infection (even bacteremia) This consensus statement is not the "last word" on
never develop sepsis, while others with seemingly these disorders. We must resign ourselves to the fact
mild infections develop a massive systemic response. that these new criteria are not perfect. In a hypothet-
As yet, we do not know wh~ This inability to identify ical example, during the study of a new therapeutic
patients at greatest risk is currently causing problems drug, the experimental treatment ofa group ofpatients
for the researchers who design clinical trials and it defined by these consensus criteria might demonstrate
will cause even greater problems as clinicians imple- that the drug is not beneficial, with the result that it
ment the innovative new treatments that are being is discarded. Yet, with the use of a different set of
developed to fight sepsis, SIRS, and MODS. criteria, the agent might be found to be beneficial. As
These new de6nitions of sepsis were developed to we learn more about the cause ofSIRS, it may become
provide maximum flexibility in classifying patients for necessary to readjust our terminolo~ Such adjust-
identification and treatment in both the clinical and ments in language would be a small price to pay if we
research settings. For instance, a patient fulfills the finally learn how to substantively improve outcome
criteria for sepsis when more than one of four criteria for patientS with these disorders.
are present; one does not have to meet a set of speci6c Roger C. Bone, M.D., F.C.C.E
and absolute criteria. For severe sepsis, sepsis must Chicago
be present with either hypotension, hypoperfusion, William) Sibbald, M.D., F.C.C.E
or organ dysfunction. Although the new definition of london, Ontario, Canada
sepsis may lead to the inclusion of patients with less Charles L. Sprung, M.D., F.C.C.E
severe disease into the category of sepsis, the defini- Jerusalem, Israel
tion entails the presence of a systemic inHammatory
response and allows a differentiation of patients with Dr Bone is Dean, Rush Medical College, and Vice President for
Medical AfFairs, Rush-Presbyterian-St Lukes Medical Center, Chi-
sepsis from those with severe sepsis and organ system ~o. Dr Sibbald is Chief, Program in Critical Care, and Professor
involvement. Such a differentiation would have a of Medicine, University ofWestem Ontario, London. Dr Sprung is
Director, Intensive Care Unit, Hadassah Hebrew University Med-
further base in different mortality rates. ical Center, Jerusalem.
New treatments, based on our understanding of the Reprint requests: Dr. Bone, Rush-Pr8sbyterian-St. LM1cei Medical
Center; 1753 wm CORgresslbrkway, Chicago 60612
molecular and cellular mechanisms that underlie the REFERENCES
systemic inHammatory response, are now being de-
1 }ocobeUJs v Ohw, 378 US 184, 197 (1964) O. Stewart concurring)
veloped and tested. It seems reasonable to assume 2 Bone RC, Fisher Cj, Clemmer n et ale A controlled clinical
that these new treatments for sepsis are most likely to trial of high-dose methylprednisolone in the treatment of severe
be effective if given as early as possible. But how do sepsis and septic shock. N Engl J Med 1987; 317:653-58
we avoid giving these agents unnecessarily to patients 3 Veterans Administration Systemic Sepsis Cooperative Study
Group. EJrect of high-dose glucocorticoid therapy on mortality
in little danger of developing SIRS or MODS? It is
in patients with clinical signs of systemic sepsis. N Engl J Med
hoped that these new definitions will eventually allow 1987; 317:659-65
clinicians to forecast which patients will develop more 4 Ziegler EJ, Fisher CJ, Sprung CL, et ale Treatment of gram-
severe forms of the disease at an earlier stage of its negative bacteremia and septic shock with HA-IA human mono-
progression. clonal antibody against endotoxin. N Engl J Moo 1991; 324:429-
36
Thus, another aim of the consensus conference was
5 Greenman RL, Schein RMH, Martin MA, et ale A controlled
to assess the various ccseverity of illness" scoring clinical trial of E5 murine monoclonal IgM antibody to endotoxin
systems for trauma and sepsis. It was the belief of the in the treatment of gram-negative sepsis. JAMA 1991; 266:1097-
conference that the use of these scoring systems-in 102

1482
6 Bone RC. Sepsis, sepsis syndrome, multi-organ failure: a plea for ment of ventilator-assisted patients at home.
comparable definitions [editorial]. Ann Intern Med 1991; 114: Realities of cost and logistics of care today require
332-33
7 Bone RC. Let's agree on terminology: definitions of sepsis. Crit
that such data be applied by practitioners in the
Care Med 1991; 19:973-76 communi~ More physicians are becoming involved
8 Sprong CL. Definitions of sepsis-have we reach~'a consenstisP with the care ofventilator-assisted patients locally near
[editorial]. Crit Care Med 1991; 19:849-51 their practice. 5 Monitoring technology now exists that
9 Canadian Multiple Organ Failure Study Group. "Sepsis"-clarity could allow continuous measurements to be done
of existing terminology-or more confusion? [editorial]. Crit Care
Med 1991; 19:996-98
using the home as a clinically suitable alternative to
the hospital provided that an appropriately designed
and managed system is available. This is desirable not
only for convenience but also because family-centered
The Management of Long-term care and monitoring at home permit involvement of
Mechanical Ventilation at Home
H orne
the most consistent observers (ventilated persons and
care for patients long-term me-
requiring family members) in a natural setting under normal
chanical ventilation will increase in the future as conditions and with the regular daily routine, which
a suitable alternative for appropriately selected can- cannot be duplicated in a facili~
didates. 1 However, after hospital discharge, clinical As the interest in home mechanical ventilation
outcomes of life-supported children and adults have grows, the need ofpracticing physicians and others for
not been well documented. In this issue (see page more rigorous research regarding vital issues such as
1500), Gilgoffet al have provided physicians and others technology assessment and long-term management
caring for ventilator-assisted persons at home with outcomes becomes more critical. Physicians do have
some dramatic observations regarding the potential some available consensus guidelines and recommen-
for medical instability depending upon pathophysiol- dations for directing their care of patients in the
o~ time of da~ and approach to technique. In this home.&-9 However, physicians continue to have major
study, clinically Significant hypoventilation (associated gaps of scientific knowledge about long-term ventila-
with apnea and seizures) was observed in patients with tion and continue to face daily inadequacies in the
high cervical spinal cord injuries who were receiving organization of home care, which are disincentives for
volume-controlled mechanical ventilation via uncuffed direct involvement in care in the home.
tracheostomy and volume-preset portable ventilator. Five years ago, research activities and agendas
The hypoventilation resulted from variable leaks regarding the scientific foundation, organization of
around the tracheostomy due to different upper airway care, and public policy research were proposed and
mechanics during sleep and wakefulness. Pressure discussed at a meeting concerning mechanical venti-
support was used to compensate for the upper airway lation in the home at the National Institutes of Health.
leaks and to achieve adequate ventilation. Since then, clinical studies such as the current obser-
The clinical observations of Gilgoff et aI provide a vations by Gilgoff et aI have been steps in the right
strong message to all physicians regarding the need direction. The value of these reported experiences
for rigorous initial and continuous evaluation of each further supports the need for those concerned about
long-term ventilated patient at home. Pressure sup- the future oflong-term mechanical ventilation at home
port and other evolving techniques and technologies to address clinical and other vital investigational issues.
(noninvasive ventilation via nasal mask)! are examples Allen I. Goldberg, M.D., F.C.C.E
of new approaches that may be suitable alternatives Chicago
for patients requiring prolonged home mechanical Secretary/freasurer, American Academy of Home Care Physicians,
ventilation. Both these newer techniques and more Chicago; Director, Section of Home Health Care, Department of
traditional methods (volume ventilation via tracheos- Pediatrics, Loyola University of Chicago, Maywood, Illfnois.
Reprint requests: Dr. Goldberg, Department of lWiatnca, Loyola
tomy) require strict selection criteria and outcome University Medical Center; Maywood, IL 60153
indicators to ensure safet}; effi~ and appropriate-
ness for the quality management ofa growing number REFERENCES
of candidates being considered for home care.
1 Goldberg AI. Mechanical ventilation and respiratory care in the
The report by Gilgoff et al comes from a respiratory home in the 1990's: some personal observations. Respir Care
rehabilitation center of excellence with decades of 1990; 35:247-59
home care experience. 3 Special regional centers can 2 Leger ~ Jennequin J, Gerard M, Robert D. Home positive
provide the components of an ideal environment for pressure ventilation via nasal mask for patients with neuromus-
the initial preparation, education, and training of the cular weakness or restrictive lung or chest wall disease. Respir
Care 1989; 34:73-9
ventilator-assisted person and family members." In 3 Goldberg AI. Home care for a better life for ventilator-
this clinically oriented research setting, observations dependent people. Chest 1983; 84:365-66
can be made which can guide the medical manage- 4 Goldberg AI. The regional approach to home care for life-

CHEST I 101 I 8 I JUNE, 1992 1483