2003 WebMD Inc. All rights reserved.

8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 1
David C. Evans, M.D. and Jonathan L. Meakins, M.D., D.Sc.
14 CLINICAL AND LABORATORY
DIAGNOSIS OF INFECTION
Surgical infection is a term that, though frequently used, is not
clearly dened. In the strictest sense, it implies infection amenable
to operative management through surgical source control, as in
the case of complicated diverticulitis or necrotizing soft tissue
infection. More generally, however, the term can refer to any infec-
tion commonly seen in surgical patients (e.g., central line infection
or postoperative pneumonia). Both denitions are pertinent, in
that the same diagnostic principles apply in each situation.
The presence of surgical infectious disease is usually deter-
mined clinically and conrmed microbiologically. Identication of
an infection is rarely incidental: most often it is sought in response
to a clinical signal. This signal is frequently fever but may be one
or more of a number of other symptoms and signs.
Most surgical infections are outpatient conditions that are easi-
ly diagnosed and treated. Infections in hospitalized patients,
whether related to the primary surgical disease or resulting from
surgical therapy, are less easily managed. The greatest challenges
in diagnosis and treatment of surgical infections arise in the peri-
operative and postoperative periods.
Terminology
Traditionally, the terms infection, sepsis, septicemia, bac-
teremia, endotoxemia, and septic shock have borne similar conno-
tations; this imprecision of terminology has led to considerable
confusion about the specic role of microbial infection as a cause
of the common clinical presentation of fever, tachycardia, and
occasional hypotension.The traditional tendency to conate these
distinct concepts in this simplistic and unrened manner has had
major implications for how antibiotics are usedor, more to the
point, misusedin surgical patients.
The human bodys physiologic response to systemic infection is
well characterized and is often referred to as sepsis. However, infec-
tion and sepsis are distinct entities.The normal septic response to in-
fection may, in fact, be completely absent in immunosuppressed pa-
tients. Most surgeons, for example, have encountered a patient
receiving high doses of steroids who has a perforated intra-abdomi-
nal viscus and fecal peritonitis but whose leukocyte count, tempera-
ture, and blood pressure are all normal. Conversely, a systemic inam-
matory response mimicking sepsis may be present in noninfected
patients.
1,2
For example, patients with acute pancreatitis, tissue
necrosis, or fractures may manifest physiologic and metabolic
changes that are indistinguishable from those associated with bac-
teremia, even in the absence of infection. Animal studies have con-
rmed that a sepsislike syndrome can occur without microbial inva-
sion of host tissues.
3
Accordingly, several clinicians have used the term sepsis syn-
drome to refer to the group of signs, symptoms, and physiologic
changes that result from a variety of sterile inammatory process-
es as well as from systemic infection.
2,4
The problem with using the
term in this way, however, is that it derives from a Greek word (sep-
sis, decay) that implies infection with microorganisms. It is there-
fore not surprising that application of the term sepsis syndrome to
noninfectious settings has led to some confusion.
5,6
To clarify the relevant terminology and provide a common ver-
nacular with which to discuss surgical infection, the American College
of Chest Physicians and the Society of Critical Care Medicine held
a joint consensus conference in 1991 that led to the publication in
1992 of the currently used terminology and denitions [see Sidebar
Denitions of Key Concepts].
7
A key outcome was the denition
of the nonspecic clinical picture of temperature, heart rate, respi-
ratory rate, and white blood cell (WBC) count abnormalities as the
systemic inammatory response syndrome (SIRS). SIRS may or
may not be due to infection; when it is, it is referred to as sepsis.When
sepsis results in organ dysfunction, the ensuing state is referred to
as severe sepsis; when it results in persistent cardiovascular decom-
pensation, the ensuing state is referred to as septic shock.
Approach to Diagnosis of Surgical Infection
Definitions of Key Concepts
Systemic inflammatory response syndrome (SIRS): This response
is manifested by the occurrence of two or more of the following
conditions as a result of infection: (a) temperature higher than
38 C (100.4 F) or lower than 36 C (96.8 F), (b) heart rate
greater than 90 beats/min, (c) respiratory rate greater than 20
breaths/min or arterial carbon dioxide tension less than 32 mm
Hg, and (d) white blood cell count greater than 12,000/mm
3
or
less than 4,000/mm
3
, or immature (band) forms accounting for
more than 10% of the neutrophils present.
Sepsis: SIRS when specifically caused by infection.
Severe sepsis: Sepsis associated with organ dysfunction, hypo-
perfusion, or hypotension. Hypoperfusion and perfusion ab-
normalities may include, but are not limited to, lactic acidosis,
oliguria, or acute alteration of mental status.
Septic shock: Sepsis with hypotension despite adequate fluid
resuscitation, with persistent perfusion abnormalities that may
include, but are not limited to, lactic acidosis, oliguria, or acute
alteration of mental status. Patients receiving inotropes or vaso-
pressors may not be hypotensive at the time perfusion abnor-
malities are measured.
Multiple organ dysfunction syndrome (MODS): MODS is the
presence of altered organ function in an acutely ill patient such
that homeostasis cannot be maintained without intervention.
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 2
Host status is assessed: normal vs. compromised/complex
Clinical signal of possible infection is noted
Assess patient characteristics and circumstances of presentation:
Health of host Intensity of physiologic response
Nature of pathogen
Cardinal signs of inflammation are
Redness Heat Pain Swelling Loss of function
Other signals include
New or persistent postoperative fever
Tachypnea/tachycardia Confusion Ileus
Response to infection in compromised or complex patients differs
from that in normally responsive patients.
Therapy is initiated
Approach to Diagnosis of Surgical Infection
Patient is evaluated for presence of infection
Begin with history and physical examination.
Perform laboratory assessment:
Obtain Gram stain and cultures of wound tissue, sputum, urine, and
drainage effluent
Consider percutaneous aspiration and microbiologic examination of
potentially infected fluid
Obtain WBC and blood chemistry measurements
Obtain chest x-ray; consider imaging of operative site
Simultaneously resuscitate, identify infectious focus, give
empirical broad-sprectrum antibiotics, and undertake
source control if able.
Antibiotic choice should reflect (1) likely source of infection,
(2) hospital- vs. community-acquired, (3) previous
antibiotic therapy.
Patient is normally responsive
Risk factors include
Advanced age Major trauma End-organ failure
Thermal injury Chemotherapy 1 chronic disease
Cardinal signs may be present, but more often, infection is
occult. Clinical manifestations may include
Confusion Ileus Gastric bleeding
Intermittent hypotension and septic shock Water retention
Delayed wound healing
Laboratory signs of occult infection include
Renal, hepatic, or respiratory failure Thrombocytosis or
thrombocytopenia Hyperglycemia and insulin resistance
Immune failure
Patient is compromised or complex
Treatment is governed by health of host, nature of response to
infection (local vs. SIRS; mild sepsis vs. septic shock/MODS),
and nature of pathogen (suspected or proven).
Withhold antibiotics until definitive diagnosis is made, unless
patient is compromised or situation is urgent.
Restore homeostasis (give fluids).
Identify and control source of infection.
Identify pathogen and give suitable antibiotics.
Patient has SIRS or uncomplicated sepsis Patient has severe sepsis or is in septic shock
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 3
Clearly, SIRS includes all the signs, symptoms, and physiologic
changes characteristic of the sepsis syndrome; however, use of the
term SIRS avoids the idea that such manifestations are necessarily
the product of infection. Sepsis may be thought of as a special case of
SIRSSIRS associated with infection [see Figure 1].The term sepsis
syndrome, although very useful in guiding clinical thinking, is insuf-
ciently precise for our current needs and probably should no longer
be used.The term septicemia should not be used either.
The crucial point is that infection and sepsis are conceptually
distinct: infection is a process, and sepsis is the response to that
process. The response provides the clinical signals that lead to
diagnosis of the initiating process. As a rule, infection, once diag-
nosed, is easily treated with antibiotics and drainage. It is the man-
agement of sepsis that is difcult [see 8:13 Multiple Organ Dysfunc-
tion Syndrome].
General Approach to Diagnosis of
Infection
The search for an infection is prompted
by a clinical signal indicating a problem in
need of resolution [see Table 1]. The signal
denotes the response of a patient to an
infectious stimulus and is a function of the
patients physiologic ability to react to the
endogenous and exogenous mediators lib-
erated through the infectious process. A
thorough history and physical examination are imperative and
should be followed by selected laboratory tests. Normally respon-
sive patients tend to show the classic signals, whereas compro-
mised or complex patients often show more subtle signs that may
only be noted as abnormalities on routine bloodwork.
The diagnostic approach to suspected infection must be modi-
ed according to patient characteristics and the circumstances of
presentation; for example, the specic differential diagnosis for
infection appearing on postoperative day 1 will clearly be different
from that for infection appearing after 1 week in the ICU. There
are three important elements at play when a surgical patient expe-
riences an infection: (1) the health of the host, (2) the intensity of
the physiologic response to infection, and (3) the nature of the
pathogen. All three factors must be considered carefully in the
diagnosis of surgical infection.
Host Status: Normally Responsive
versus Compromised or Complex
That signs and symptoms of infection in
compromised or complex patients differ
from those in normally responsive hosts
has important diagnostic implications.
Normally responsive patients, for whom
the physician can obtain a history and per-
form a physical examination, respond to
infection in the classic mannertypically,
with fever, tachycardia, leukocytosis, malaise, and other appropri-
ate symptoms. For many reasonseven simply if the infection is
severenormally responsive patients may become compromised.
Compromised or complex patients are unable to meet inamma-
tory or infectious challenges in the normal manner.
Hence, the clinical signals of infection
in such patients differ from those in normally responsive patients,
often being absent or developing at a later stage of infection.
Indeed, a multitude of clinical conditions or physiologic states
dene compromised or complex patients. These include the ex-
tremes of age, immunosuppression as a result of either disease
(e.g., HIV infection or lymphoma) or medication (e.g., chemo-
therapy), thermal injury, major trauma, acute end-organ failure in
the ICU, and the presence of more than one chronic disease. The
prevalence of such patients in modern hospital surgical practice is
increasing steadily.
It must be kept in mind, however, that the normally responsive
patient and the compromised or complex patient are merely
extreme points on the clinical spectrum rather than categorically
distinct populations.
Physiologic Response to Infection
NORMALLY RESPONSIVE PATIENTS
Cardinal Signs of Inammation
Rubor, calor, dolor, tumor, and functio
laesathat is, redness, heat, pain, swelling,
and loss of functionhave been consid-
ered the cardinal signs of localized inam-
mation since the times of Hippocrates and
Galen. They remain the primary signals
leading to medical consultation for outpatient surgical infections
and for many of the infectious complications of operation. They
are emblematic of the hosts effort to contain infection locally and
may signal the presence of infection even in cases where the pri-
mary site of infection is a deeply situated organ or tissue.
Infections amenable to surgical intervention may present in this
way; however, nosocomial infections complicating the course of
surgical patients are generally signaled in more subtle ways.
Fever
Fever is perhaps the most common signal that an infectious
process is present.
Postoperative fever is a normal part of the recovery process; under-
standing the typical febrile course is important in differentiating
normal from pathologic fever. It is unusual for a sudden, very high
fever to be the rst signal of an infection. Infection usually begins
to manifest itself with a prodrome, recognition of which speeds
diagnosis and therapy. Investigation should be started when the
patients temperature reaches 38 C (100.4 F) rather than 40 C
(104 F). Although this point may seem obvious, many of the cri-
SIRS Infection
Bacteremia
Fungemia
Parasitemia
Viremia
Other
Trauma
Other
Burns
Pancreatitis
Sepsis
Figure 1 Depicted are the interrelationships among infection, sep-
sis, and the systemic inammatory response syndrome (SIRS).
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 4
sis intervention measures required in managing fevers could be
avoided if the signicance of more modest temperature elevations
were recognized more often.
A fever that appears after the normal postoperative temperature
elevation has resolved must not be ignored.To simply wait for such
a fever to dissipate is to court disaster. In the absence of a clear
diagnosis, a thorough physical examination of the patient, direct-
ed by laboratory tests and followed by reexamination as necessary,
is required to identify occult infection.
Miscellaneous Signals
It is common wisdom that the signals communicating under-
lying infection in the compromised or complex host may be sub-
tle.The astute clinician will be in tune with these and, with expe-
rience, will recognize when to undertake a diligent search for
infection.
Altered heart rate A heart rate that is either too high or too
low may signal an infection. On rare occasions, a change in rhythm
in the elderly (e.g., paroxysmal atrial tachycardia, utter, or atrial
brillation) indicates an infectious process. Gram-negative sepsis
may produce a so-called relative bradycardia, meaning that the
resulting tachycardia is not as pronounced as one might expect.
An unexplained sustained increase in heart rate should not be
ignored.
Tachypnea Whereas tachypnea occurs commonly after oper-
ation in response to pain or poor pulmonary toilet, it may also sig-
nal either the prodrome of infection or the onset of SIRS. Because
tachypnea may herald not only infection but also other important
diagnoses (e.g., pulmonary embolism), it must be thoughtfully
and methodically evaluated.
Pain Pain that persists or is out of proportion to the expect-
ed response deserves attention. Whenever a surgical wound that
was healing favorably for the rst 5 to 7 days becomes more
painful, a deep surgical site infection (SSI) must be suspected and
ruled out, even if other signs are absent. Unexplained muscular
pain is often the rst harbinger of deadly necrotizing soft tissue
infection caused by gram-positive bacteria (e.g., group A strepto-
cocci), the early recognition of which may be lifesaving and limb-
preserving [see 3:2 Soft Tissue Infection]. Sometimes pain is re-
ferred, and the painful area appears normal on examination. Pneu-
monia that presents with abdominal ndings is a classic example,
as is the shoulder-tip pain with a normal range of motion seen in
patients with a subphrenic abscess.
Confusion Confusion is a common symptom of infection in
the elderly; it is also an important signal in patients who had been
well and t.The physicians rst response to confusion in an elder-
ly patient in the postoperative period must be to seek a cause, not
to order sedation.
Ileus Ileus has many causes, some of which are not well
understood. Prolonged ileus after abdominal operationas well as
almost any ileus after other operationsrequires explanation.
Infections at remote sites (e.g., SSI and pneumonia) can produce
ileus, as if the bowel were a target organ such as the kidney, the
liver, or the lung.
COMPROMISED OR COMPLEX PATIENTS
The presence of fever remains a common
signal of infection in the compromised or
complex patient; however, it may be absent,
or the patients temperature may already be
elevated as a result of other causes. Cardinal
signs of infection may be present. More of-
ten, however, the infection is occult, and
classic signals are unrelated to the infectious
focus. In some very ill immunocompro-
mised patients, ndings that usually signal an infection may already
be present. Slight changes in clinical status (e.g., minor temperature
elevations, increased uid requirements, confusion, and ileus) or
changes in laboratory ndings (e.g., an elevated WBC count, gluco-
suria, and hyperglycemia) should trigger investigation.
Patients in whom the rst signal of an infectious process is organ
dysfunction or failure, rather than fever and tachycardia, are likely to
be physiologically compromised and seriously ill; perhaps more im-
portant, however, is that they are a group whose diagnosis and man-
agement require expert clinical skills. Because the classic septic re-
sponse may not be present, it is essential to be alert to the signs and
symptoms of occult infection (see below). In these patients, the labo-
ratory and the radiology suite become increasingly important in di-
agnosing and documenting the evolution of the infection.
Clinical Signs and Symptoms of Occult Infection
Subtle changes in temperature, mental status, pulse rate, or res-
piratory rate may signal occult infection, as may the development
of pain or ileus.
Intermittent hypotension and septic shock Septic shock,
an important manifestation of an unrecognized focus of infection,
is the original expression of multisystem failure. Fortunately,
it rarely occurs without warning. The prodrome often includes
fever and sometimes other signals. Recurring hypotension is the
most characteristic signal; it usually is not catastrophic and
responds quickly to uid resuscitation. Oliguria may accompany
the hypotension. If this clinical state is allowed to progress, the
hypotension will lead to renal failure (see below), with substantial
water retention. Septic shock will result if the infection is not iden-
tied and treated with appropriate antibiotics and source control
as necessary.
Both clinical assessment and laboratory studies are necessary
to conrm the presence of septic shock, although orid septic
shock is easily recognized on clinical examination alone [see 8:3
Table 1 Fundamental Approach to
Diagnosis of Infection
Recognize clinical signal and observe its characteristics:
Nature
Intensity
Rapidity of development
Make best guess as to source and likely pathogen on the basis of
History of surgical disease
Physical examination
Microbiologic examination of stained specimens
Radiologic findings
Confirm presence of infection by means of
Laboratory results
Observation of clinical course
Invasive procedures (e.g., paracentesis, thoracentesis,
interventional radiology, operation)
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 5
Shock]. Any or all of the following ndings may be present in vary-
ing degrees: tachycardia; tachypnea; hypotension; warm, dry ex-
tremities; generalized ushing; and other signs suggesting a hyper-
dynamic, hypermetabolic state. Swan-Ganz catheter measure-
ments conrm high cardiac output and low peripheral vascular
resistance.
Gastric hemorrhage Gastric hemorrhage may be the present-
ing symptom of serious infection even if prophylactic measures
against such hemorrhage have been taken. It is particularly sugges-
tive of perigastric abscess resulting from anastomotic leakage after
upper abdominal surgery. Gastric hyperacidity and bleeding general-
ly respond to drainage of an abscess. Hemorrhagic gastritis must al-
ways be considered a signal of occult infection, which demands
prompt diagnosis and treatment.
Delayed wound healing The absence of wound healing can
indicate the presence of a signicant infection. Typically in such a
case, wounds left for delayed primary closure or secondary closure
do not exhibit the appropriate granulation tissue and appear pale,
dry, and unhealthy.The development of good granulation tissue is
a sign that infection is controlled.
Laboratory Signs of Occult Infection
Renal failure Renal failure [see 8:6 Renal Failure] is identi-
ed by elevations in serum creatinine and blood urea nitrogen
(BUN) levels, which can be highly sensitive signals of developing
infection. A still more sensitive indicator is an alteration in creati-
nine clearance, a laboratory test underutilized in the ICU. Such
alterations are generally evident before changes in serum levels.
Creatinine clearance should be measured at an early stage in high-
risk patients. In the presence of shock, renal failure can develop
suddenly. Otherwise, loss of renal function is insidious, but it can
usually be identied if sought before oliguria or anuria develops.
Resolution of infection is associated with return of function.
Hepatic and respiratory failure Hepatic failure [see 8:9
Hepatic Failure], primarily manifested as jaundice, and respiratory
failure, initially presenting as a
falling arterial oxygen tension and subsequently marked by a need
for mechanical ventilation or by a change in the fraction of
inspired oxygen (F
I
O
2
) requirement, can behave in the same way
as renal failure (i.e., with drainage and control of infection leading
to restoration of function).
Abnormal platelet count Thrombocytosis is often seen in
association with infection, particularly with compromised hosts, in
whom the infection may be occult. Thrombocytopenia may also
indicate serious infection, though it is not a common signal.When
this occurs in the context of sepsis, it is either because disseminat-
ed intravascular coagulation (DIC) has developed or because
there is a diminution of all blood lines indicating marrow dys-
function. The cause of any abnormal platelet count should be
identied promptly if possible.
Hyperglycemia and insulin resistance Hyperglycemia
and insulin resistance are often reliable signals of the presence of
infection in diabetic patients as well as in nondiabetic patients.The
degree of insulin resistance can reect the severity of the infection
as well as the effectiveness of infection control.
Immune failure Immune failure is discussed in detail else-
where. Severe infection is
immunosuppressive. The most clinically applicable measure of
immune failure at present is probably delayed wound healing.
Evaluation for Presence of Infection
Once the surgeon has evaluated in
which ways the patient may be compro-
mised or susceptible to infection and to
which degree the physiologic response is
inappropriate or harmful (i.e., provokes
multiple organ dysfunction or shock), then
it is important to consider which microor-
ganisms might be responsible for the pre-
senting clinical picture.This obviously varies
with the clinical situation and requires knowledge of specic con-
dition-associated pathogens, as well as the prevalence of certain
pathogens in a particular hospital or ICU.
The prevalence of resistant bacterial strains should be moni-
tored in clinical settings in order to guide empirical therapy.
Methicillin-resistant Staphylococcus aureus (MRSA), once an occa-
sional nding, has become so common in some ICUs and chron-
ic wards that surveillance and isolation programs are no longer
employed. It is now the rule for extensive invasive monitoring and
access devices to be used in the care of critically ill patients, who
thereby become particularly predisposed to gram-positive infec-
tion; accordingly, it is important to appreciate the likelihood that
MRSA will be encountered. The same is true of Pseudomonas
aeruginosa, a ubiquitous commensal and a common gram-nega-
tive pathogen in hospitals. In compromised patients exposed to
multiple antibiotics, P. aeruginosa readily acquires antibiotic resis-
tance that usually necessitates the use of double-agent or broad-
spectrum coverage for effective management.
Recognizing the virulence of certain pathogens is as important
as appreciating their antibiotic susceptibilities. Enteroinvasive
Escherichia coli 0157-H7, for example, may cause a rapidly pro-
gressive hemorrhagic enteritis and provoke a fatal septic syndrome
marked by acute renal failure, bleeding, and coma. Necrotizing
soft tissue infections, particularly when caused by gram-positive
organisms, may be precipitously fatal. Early Gram stain microscopy
to identify the specic pathogen is a critical step in the manage-
ment of this condition.
Essentially the same clinical and laboratory assessments are
used to evaluate normally responsive and compromised or com-
plex patients for the presence of infection.There is, however, a sig-
nicant difference in emphasis. In normally responsive patients,
the diagnosis of infection is usually made on clinical grounds with
laboratory support, whereas in compromised or complex patients,
the diagnosis is usually made on the basis of laboratory ndings
with clinical support.
The ICU patient presents a particular conundrum. Nosocomial
infection is identied in an estimated 20% of such patients.
8
Despite the frequency with which it is suspected and reported, it
is difcult to prove unequivocally. The perceived prevalence of
nosocomial infection has created a strong predisposition toward
instituting empirical antibiotic therapy in ICU patients; however,
the global value of this action in terms of both patient outcome
and the impact on the ecology of the ICU is unconrmed and
requires validation. The enormous inconsistencies in how infec-
tions are diagnosed have a tremendous effect on our ability to
assess the efcacy of therapy for infection. The current approach
to diagnosing infection in surgical patients, particularly those who
are critically ill or compromised, is still in great need of clarica-
tion and standardization.
9
Until these issues are resolved, the clin-
ician must be familiar with the strengths and limitations of a vari-
ety of current diagnostic methodologies and then exercise
thoughtfulness and disciplined diligence.The likelihood that a par-
ticular patient is infected (i.e., the pretest probability) is as impor-
tant in the decision to treat as the fulllment of any particular con-
stellation of diagnostic criteria.
HISTORY AND PHYSICAL EXAMINATION
The history should include all comorbid conditions (e.g., dia-
betes, lung disease, cirrhosis, hepatitis, kidney disease necessitating
dialysis, and previous important infections) as well as a hospital-
ization history that covers health status, surgical diagnosis and
therapy, additional therapeutic interventions (including interven-
tional radiology, monitoring devices, drains, and drugs), and other
related variables.
In the early postoperative period (3 to 6 days after operation),
the traditional causes of the signals of infection have their origin in
the wound, intravascular lines, the urinary tract, and the lungs.
Deep thrombophlebitis, with or without pulmonary embolism,
may also initiate a systemic inammatory response that mimics
sepsis.The general physical examination is often unrewarding, but
a number of specic examinations should be carefully performed,
with emphasis given to (1) all wounds and surgical sites, (2) all
invasive monitoring or therapeutic devices and surrounding areas
(notably central and peripheral I.V. lines), (3) all drainage systems
and surrounding tissue, with particular attention paid to the
nature of the drainage and whether it has recently changed in
character or volume (particularly if it has stopped), (4) the rectal
examination (for pelvic or prostatic infection), (5) areas of poten-
tial decubitus ulcers, (6) the neck (for CNS infection), (7)
intravascular lines, surrounding tissue, and proximal vessels, (8)
the lungs, and (9) the legs. The physical examination is important
as a guide for selecting specimens for microbiologic analysis, par-
ticularly when there have recently been signicant changes in
wounds or drainage. The decision to seek radiologic consultation
may depend on the ndings on physical examination.
DIAGNOSTIC TESTS
Hematologic and Biochemical Tests
After physical assessment, laboratory blood tests are routinely
relied on to orient the surgeon toward or away from a clinical
diagnosis of infection. Leukocytosis, particularly with an
increase in band forms, is a usual but inconsistent marker for
infection. The WBC count is widely used to follow the response
of infection to therapy and thus has been adopted as a surrogate
indicator of the success or failure of therapy. Surprisingly, how-
ever, the documentation supporting this ubiquitous practice is
sparse.
10,11
Not only is the daily series of complete blood counts
often ordered in conjunction with the initiation of antibiotic
therapy wasteful and unpleasant for the patient, but it also typi-
cally tells the clinician little about the clinical course that cannot
be gleaned at the bedside.
In more complex surgical patients, other biochemical cues are
used to varying degrees as means of assessing the likelihood of infec-
tion. In addition to thrombocytosis, thrombocytopenia, hyper-
glycemia, and metabolic acidosis, which commonly reect the stress
of severe infection, changes in the erythrocyte sedimentation rate
(ESR) and in blood levels of C-reactive protein (CRP), procalcitonin
(PCT), interleukin-6 (IL-6), and tumor necrosis factor (TNF) have
a signicant association with the presence of clinical infection. Plas-
ma CRP concentration, which has been extensively used in some
European countries to monitor the evolution of infection, has been
found to be signicantly elevated in patients with pulmonary aspira-
tion that has induced bacterial infection, compared with patients
with sterile pneumonitis.
12
Some investigators have suggested that
because CRP concentration appears to be particularly responsive to
bacterial infection, it may be useful as a monitor of the efcacy of an-
tibiotic use, thereby guiding discontinuance of treatment.
13
A host of
cytokines, cellular adhesion molecules, oxidants, and other biomole-
cules known to participate in systemic inammation from numerous
causes are being extensively investigated to establish both diagnostic
and therapeutic functions [see 8:26 Molecular and Cellular Mediators
of the Inammatory Response]. As yet, however, no single mediator of
systemic inammation has been validated as a reliable clinical tool
for surveillance of the progression of infection or the response of in-
fection to treatment.
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 6
Infection is a process; sepsis is the response.
The difficulty of managing the patient who manifests the septic re-
sponse in the absence of infection is illustrated by the following case.
A 55-year-old insulin-dependent man with peripheral vascular
disease presented with evidence of infection in both feet.
Hydration and antibiotic therapy did not prevent progression of the
infection, and within 18 hours, it was apparent that amputation
would be required for source control. In the course of the opera-
tion, gas gangrene, more extensive than had been clinically sus-
pected, was discovered. The initial below-the-knee amputations
were eventually followed by a hip disarticulation on one side and a
high above-the-knee amputation on the other. Over the 36-hour
period during which the infectious process was being controlled,
classic septic shock, renal failure, coma, and respiratory failure
developed. There was no change in the patients hyperdynamic
and hypermetabolic state after the amputations. During the next
3 weeks, he required ventilator support and daily hemodialysis or
hemofiltration; became jaundiced and more deeply comatose;
received fluids in amounts significantly in excess of output to main-
tain blood volume; was hyperglycemic despite receiving regular
insulin in dosages of 3 to 5 U/hr; and remained in a hyperdynamic
state. Shortly after the last operation, an ileus developed, accom-
panied by gross fluid retention, which further increased the
patients girth. This state of overt sepsis with hypermetabolism per-
sisted while the wounds healed by primary closure, but no focus of
infection could be found.
Antibiotic therapy was stopped for 10 days after operation; the
patients clinical status did not change. Frequent searches were
made to ensure that no infection had been overlooked. Sug-
gestionsseriously put forwardto explore the patients abdomen
because there must be something there were not heeded. At the
end of the third week, for no obvious reason, the patient started to
urinate, his ileus resolved, and he was gradually weaned from
the ventilator. His level of consciousness improved, the massive
edema cleared, and the high cardiac output and low peripheral
resistance resolved over a period of 72 hours. He was discharged
from the surgical ICU 1 day later and from the hospital in 3 weeks.
Some noninfectious process that had maintained this patients per-
sistent septic response had disappeared or had been turned off,
and the result was rapid resolution of the septic state and recovery
of health.
As noted, initial control of infection, though difficult, was
achieved relatively early. Subsequent therapeutic efforts involved
providing hemodynamic, metabolic, and physiologic support of
the patients failing organs and organ systems while waiting for the
septic response to resolve. The real problem in this case was not
the infection but the patients unremitting septic response, which
was initiated by the infection but maintained in its absence.
Case Study: Clinical Picture of Sepsis without
Infection
Microbiologic Studies
As a rule, Gram stains and cultures of wound tissue, sputum,
urine, and drainage efuent are useful studies. In some cases, a bat-
tery of cultures of potential sites of infection may be the only feasible
approach. Culture techniques are discussed more thoroughly else-
where [see 8:27 Blood Cultures and Infection in the Patient with the Sep-
tic Response and 8:16 Nosocomial Infection].
The use of polymerase chain reaction (PCR) technology to detect
bacterial DNA is emerging as a useful alternative to microbiologic
culture for determining the presence of infection. PCR identies and
amplies a specic bacterial DNA sequence by means of a chemical
proliferation process that may take no longer than a few hours. Un-
fortunately, the sensitivity and specicity of this powerful technique
is highly variable. Although investigations have shown PCR to be a
sensitive method of conrming the presence of bacteria in the blood
of clinically septic ICU patients,
14
comparison of PCR results with
blood culture results is problematic in that nonviable bacteria or bac-
terial debris are likely to create false positive results, leaving the clini-
cian uncertain as to whether treatment is indicated.The accuracy of
PCR thus remains to be established, but the rapidity with which it
yields results makes it highly promising as a potential guide to thera-
peutic intervention.
Radiology
A chest x-ray is mandatory. Ultrasonographic examination of the
operative site may be useful to evaluate the possibility of a deep ab-
scess. Computed tomography of the operative site is often more use-
ful than ultrasonography because the presence of wounds, dressings,
and drainage tubes may obscure ultrasonographic ndings.The pos-
sibility of acalculous cholecystitis must be kept in mind, though this
is probably an overdiagnosed entity of unvalidated clinical signi-
cance.
In compromised or complex patients who have
not recently undergone an operation, the medical and surgical histo-
ry combined with the radiologic examination may be the only guide
to the potential infectious focus (e.g., ulcer, diverticulitis, cholecysti-
tis or cholangitis, or obstructed ureter). Percutaneous aspiration of
potentially infected uid should be considered, and this uid should
be microbiologically examined if possible.
Institution of Therapy
As noted, the response to infection occu-
pies a continuum ranging from virtually no
clinical expression in the immunosup-
pressed patient to full-blown septic shock.
Sepsis, in its basic form, represents a mild
to moderate response to infection that
occurs in normally responsive patients as
well as in many compromised or complex
patients. On the basis of the results of clin-
ical or laboratory tests, the clinician can evaluate the magnitude of
the septic response and thus its clinical gravity.
It is therefore the magnitude of the host response, in combina-
tion with the health of the patient and the nature of the likely
infecting organism, that directs the clinicians approach to therapy.
The greater the degree of sepsis, the greater the clinical urgency of
solving the two fundamental problems involved: resolution of the
initiating process (i.e., treatment of the infection) and modulation
of the response to that process (i.e., management of sepsis) [see
8:13 Multiple Organ Dysfunction Syndrome]. The therapeutic
approach to a given patient is based on the need for speed.
SIRS
As mentioned, SIRS was specically dened in such a way as to
emphasize that this common clinical syndrome frequently occurs
without infection and does not routinely necessitate treatment,
especially with empirical antibiotic therapy. As an illustration, sim-
ply climbing a set of stairs is often enough to enable a healthy indi-
vidual to fulll the diagnostic criteria for SIRSalbeit briey.
The point that SIRS is not invariably associated with microbial
invasion is especially relevant to critically ill and traumatized ICU
patients, in whom it is not uncommon for clinical signs and symp-
toms indistinguishable from those of severe sepsis to arise or per-
sist in the absence of any infection [see Sidebar Case Study:
Clinical Picture of Sepsis without Infection]. Burn injury and pan-
creatitis are classic examples of conditions that can provoke such
a response: both can give rise to a hyperdynamic, hypermetabolic
clinical picture identical to that of sepsis or severe sepsis, even
when no infection is present. Surgeons have learned not to give
antibiotics unless there is evidence of infection. The instinctive
reex to do something must be held in check: masterful inactivi-
ty is the appropriate response until a specic source that can be
controlled is identied.
UNCOMPLICATED SEPSIS
The approach to treatment of infection in both normally respon-
sive and compromised patients with mild to moderate sepsis in-
cludes ve steps: (1) resuscitation and reestablishment of homeosta-
sis and organ function, if necessary; (2) identication of the focus of
infection by clinical or radiologic examination; (3) source control,
which implies removal, containment or control of the infectious fo-
cus (e.g., open drainage of an infected wound; resection of compro-
mised bowel, as for appendicitis or advanced diverticulitis; or percu-
taneous drainage, as for pancreatic abscess); (4) microbiologic
characterization of the offending pathogen by means of culture,
Gram stain, or both; and (5) empirical or targeted treatment with
antibiotics, depending on the clinical urgency.
In compromised patients, empirical antibiotic therapy is often
started before diagnosis, and its efcacy gauged by the patients
subsequent clinical course [see Discussion, below]. In normal hosts
with uncomplicated sepsis, many would consider it appropriate to
withhold empirical therapy until source control has been effected
and the pathogen or pathogens characterized. Adequate source
control, when possible, is often all that is required for successful
management of sepsis, particularly in healthy hosts. In such cases,
antibiotic use for prophylaxis against complications such as SSIs is
of proven benet; however, there actually is little evidence to indi-
cate that antibiotics are necessary to treat sepsis if mechanical
source control has been denitively achieved.
Antibiotic Treatment
In the absence of convincing culture data, antibiotic treatment
must be directed against a likely cause, as determined by recent
history (particularly procedures), past history, and physical exam-
ination. Examples of likely causes are (1) urinary manipulation,
which necessitates coverage against enterococci and gram-nega-
tive bacteria with ampicillin and an aminoglycoside; (2) colonic
ora, which mandates wide coverage against anaerobes and aer-
obes; (3) infected vascular lines, which warrant coverage against
gram-positive organisms; (4) cholangitis, which calls for coverage
against aerobic gram-negative bacteria with ceftriaxone or an
aminoglycoside; and (5) pneumonia, which necessitates coverage
against gram-positive and gram-negative aerobes.
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 7
Drainage
The search for the focus of infection is important because
drainage may resolve the entire problem. The clinical state can
be changed dramatically by technical or mechanical manage-
ment of pus behind an obstruction (e.g., in the biliary tree, the
urinary tract, or the tracheobronchial tree) or pus under pres-
sure (e.g., an abscess), by manipulation of a drain, or by
removal of a foreign body (e.g., an intravascular line). Prompt
elimination of all foci of infection that can be drained or are
operable is critical. Needle aspiration of peritoneal or pleural
uid may be very helpful. Wounds must be reevaluated con-
stantly and the presence of pressure sores ruled out. Drainage
can often be performed either percutaneously (e.g., for
pyelonephritis from ureteral obstruction or for subphrenic
abscess) or endoscopically (e.g., for cholangitis). Antibiotics
alone may sufce to treat small collections not accessible by
means of percutaneous techniques.
SEVERE SEPSIS AND SEPTIC SHOCK
For assessment and management of se-
vere sepsis and septic shock [see Sidebar
Denitions of Key Concepts], the four
steps in treatmentresuscitation, diagno-
sis of the infectious focus, antibiotic thera-
py, and drainagemust be performed
concurrently. Specically, I.V. administra-
tion of empirical antibiotics must be initi-
ated promptly, before the diagnostic
process is completed, and any potentially drainable focus must be
identied via physical examination or radiologic imaging. The
choice of antibiotic depends on (1) what the likely source of infec-
tion is (e.g., a lung, a perforated viscus, or the biliary tract), (2)
whether the infection is hospital acquired (in which case antibiot-
ic resistance must be considered), and (3) whether the patient has
previously received antibiotic therapy.
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 8
Discussion
Approaches to Specic Infections in Complex Surgical
Patients
The following infections may occur in all surgical patients.They
are, however, much more difcult to identify in complex surgical
patients, such as those admitted to the ICU. Because the signals of
infection are less specic and extremely difcult to interpret after
injury or operation, the approach to diagnosing and treating infec-
tion must be cautious and disciplined.
SURGICAL SITE INFECTION
SSIs include all infections occurring within the operative eld,
from the skin to the actual area of surgery [see 1:1 Prevention of
Postoperative Infection].
15
The patient history should address previ-
ous diseases as well as issues concerning the operation itself, such
as wound risk classication, duration, difculty, urgency, use of
drains, other details of the procedure, time elapsed since the oper-
ation, and whether the patient was immunosuppressed, experi-
enced trauma, or received chemotherapy. Physical examination
should focus on the cardinal signs of infection and the absence or
presence of a healing ridge.
Deeper infections tend to become apparent later in the postop-
erative course, often after a period in which the patient appears to
be recovering, and are associated with a variety of signals, some of
which can appear suddenly.The physical examination is often use-
less or misleading because of discomfort associated with the oper-
ation. Surgical site pain that increases or fails to resolve in the 7 to
10 days following surgery is an important yet subtle marker for
occult infection that calls for investigation. Rectal examination is
important because it may detect abscess formation or bleeding.
Return of ileus after an abdominal operation is a signicant clue to
the presence of abdominal infection.
Culture is essential because use of the correct antibiotics is par-
ticularly vital in treatment of compromised or complex patients.
Knowledge of the organism and its sensitivities is the key to iden-
tifying epidemic or multiresistant strains.
URINARY TRACT INFECTION
Nearly all patients admitted to the ICU have a urinary catheter
in place; of these, it is estimated that about 20% progress to uri-
nary tract infection (UTI).
8
Bacteria adhere to urinary catheter
surfaces, where they promote growth of a so-called biolm com-
posed of microorganisms, bacterial glycocalices, Tamm-Horsfall
protein, and urinary salts. Eradication of this infectious nidus is
essentially impossible without catheter removal.
16
The standard
criterion for the diagnosis of UTI (10
5
bacteria/ml) is difcult to
apply in catheterized patients because antibiotic therapy without
removal of the catheter and the source of bacteriuria would be
ineffective. Furthermore, it is well established that urine cultures
demonstrating as few as 10
1
or 10
2
bacteria/ml increase 1,000-fold
within 1 to 2 days
17
; effectively, therefore, any bacterial growth on
a urine culture from a catheterized patient signals heavy coloniza-
tion, if not infection.
More important than quantifying the degree of bacteriuria is
determining whether there is any evidence of tissue invasion by
urinary bacteria, which would present as pyelonephritis, cystitis,
prostatitis, epididymitis, bacteremia, or sepsis. The patient history
should determine whether a Foley catheter was used and for how
long; how, when, and why it was inserted; instrumentation (e.g., a
so-called in-and-out catheter, cystoscopy, or transurethral resec-
tion of the prostate); and whether the patient has had any previous
UTIs. The physical examination should ascertain whether there is
any costovertebral angle tenderness or evidence of prostatic or epi-
didymal tenderness.
Laboratory tests should include gross and microscopic urinaly-
sis, urine culture, and sensitivity tests. Blood culture is important
because it may substantiate the diagnosis, identify the bacteria pres-
ent, and determine the degree of invasiveness of the infectious
process.
VASCULAR CATHETER INFECTION
The most frequent sites of infection postoperatively are I.V.
catheters, particularly peripheral ones. Diagnosis of peripheral
catheter infection is simple and is made on clinical grounds.
Diagnosis of central venous catheter (CVC) infection is more dif-
cult. Because hospitalized patients are increasingly being man-
aged with monitoring or therapeutic modalities that depend on
vascular access (e.g., total parenteral nutrition and dialysis), line
infections have become more common, with an incidence ranging
from two to 30 infections per 1,000 CVC days.
18
The combined
pressures imposed by (1) the need to maintain vascular access in
sicker and more complex patients and (2) the increasing predom-
inance of gram-positive CVC infections observed since the late
1970s have led clinicians in many centers to administer empirical
therapy without line removal to complex patients as a matter of
course; some even advocate 10 to 21 days of vancomycin-based
therapy. The problems associated with the latter approach
emerging vancomycin resistance, nephrotoxicity, and rashare
serious and relate specically to the diagnostic strategy used to
manage potential CVC infections. Distinguishing between contam-
ination, colonization, and true infection is problematic; as a result,
a number of diagnostic strategies have been advocated that are
predicated more on practicality and cost-effectiveness than on
microbiologic reality.
It is believed that CVC infection most commonly arises from
invasion by skin microorganisms (S. aureus or S. epidermidis in
about 80% of cases), which may manifest itself as exit-site puru-
lence with or without local cellulitis, as a tunnel infection that may
be clinically difcult to detect, or as catheter-related bloodstream
infection (CR-BSI). Of these, CR-BSI, which complicates as many
as 5% of line placements, is the most clinically important entity. It
is strictly diagnosed by identication of the same microorganism
(identical species and antibiogram) grown from both the catheter
and a peripheral blood culture.
The catheter may be cultured in one of several ways, the most
common of which is the roll-plate method [see 8:16 Nosocomial In-
fection]. Because it is theoretically possible that this technique may fail
to detect bacteria harbored within the catheter lumen, some authori-
ties advocate the more sensitive sonication method, in which the
catheter segment is immersed and agitated in a medium to produce a
broth that contains bacteria from both the internal and the external
surfaces of the line.This technique is both more costly and more time
consuming, in that it requires quantitative cultures that are deemed
positive only when more than 10
3
colony-forming units are detected.
More often, blood drawn through the CVC or cultured from an exit-
site exudate is compared with peripheral cultures, and thus there is
no need to remove the line. If quantitative cultures are done, a line
blood culture showing ve to 10 times more growth than the periph-
eral sample strongly suggests that the catheter is the source of the
bacteremia. A less costly method that renders quantitative cultures
unnecessary relies on the speed of bacterial growth: if growth in
catheter-drawn blood is faster than that in peripherally drawn blood,
a primary line infection is likely. On its own, a line blood culture is
not sensitive or specic enough to be diagnostically useful.
PULMONARY INFECTION
Diagnosis and management of nosocomial pneumonia in surgical
patients is addressed in detail elsewhere,
The central issue is that there is no universal agreement
as to how pneumoniaparticularly ventilator-associated pneumo-
niashould be diagnosed. Of the innumerable diagnostic options,
none can rely on demonstration of tissue invasion by microorgan-
isms, as would be ideal: all are to some degree nonspecic, and any
may be invoked to justify initiation of antibiotic therapy.
19
Random-
ized trials linking mode of diagnosis to therapeutic strategy and then
to outcome are yet to resolve this issue.
20
SINUSITIS
All patients undergoing prolonged nasogastric intubation are
predisposed to sinus infection. Previous facial trauma and a histo-
ry of sinusitis are potential contributing factors as well. (Otitis and
pharyngitis, which are not often considered, occur in much the
same group of patients.) Because maxillary or frontal area tender-
ness is nonspecic in very ill surgical patients, the diagnosis is usu-
ally based on CT demonstration of sinus opacication or air-uid
levels. As a rst step, sinus drainage should be reestablished by
removal of an unnecessary nasogastric tube. Some authorities
advocate sinus aspiration for culture before empirical antibiotic
therapy is begun in urgent cases.
PAROTITIS
Parotitis is an increasingly common clinical diagnosis in elderly
patients. It is usually caused by S. aureus and diagnosed on the
basis of the presence of the classic local signs of inammation.
Culture of Stensens duct and blood culture are useful.
PROSTATITIS
Prostatitis (diagnosed by rectal examination) and epididymitis
are clinical expressions of Foley catheterrelated infection.The aid
of prostatic massage is important in obtaining specimens for cul-
ture. It should be remembered that a blocked Foley catheter is the
most common cause of hospital anuria. This obstruction can lead
to devastating purulent cystitis and upper UTI.
PSEUDOMEMBRANOUS ENTEROCOLITIS
Antibiotic-associated pseudomembranous enterocolitis is diag-
nosed by obtaining specimens for serology and culture and per-
forming proctosigmoidoscopy. Clostridium difcile is frequently
identied as the causative pathogen. Although pseudomembran-
ous colitis is rarely clinically impressive and is easy to overlook, it
can be rapidly fatal. Initial appropriate antibiotic therapy is not
always successful; therefore, reevaluation at the end of the treat-
ment course is required.
Problems with Empirical Treatment of Infection
The frequent presence of SIRS in complex or critically ill surgi-
cal patients usually prompts a reexive response to pan-culture
the patient if no credible source of infection is apparent.When per-
missive or loose diagnostic criteria for infection are invoked, the
inevitable result is the commencement of empirical antibiotic ther-
apy for suspected infection, which is often, by default, continued
for days, if not weeks, pending denitive culture results or clinical
improvement. This strategy may seem reasonable and is under-
standably difcult to resist, but it is potentially deleterious in many
ways, and there are many sound objections to its reexive use.
21
Empirical antibiotic therapy can obfuscate future cultures, pre-
dispose to the emergence of resistant organisms (which are associ-
ated with increased attributable mortality), promote derepression
of homeostasis-maintaining endogenous ora, cause toxic reac-
tions and secondary effects, alter the ecology of the unit in which
it is used (as shown by the rising prevalence of MRSA and van-
comycin-resistant enterococci in both European and North
American centers), and raise the cost of patient care. This widely
used strategy is largely unvalidated.
10,11
It must be emphasized that
the paramount principle of therapy for infection is treatment
focused on appropriate microbiologic cultures in the context of
strict diagnostic criteria for infection.
Many authorities espouse streamlining of empirically begun
broad-spectrum antibiotic therapy in response to microbiologic
data once culture results are available; however, it is frequently dif-
cult to discontinue antibiotics once they have been started. When
strict diagnostic criteria for infection are not met or, more impor-
tant, when antibiotic therapy based on nonmicrobiologic evidence
of infection yields negligible results, strong consideration should be
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 9
given to stopping the antibiotics, and an exhaustive effort should be
made to identify and control any occult persistent cause of the
inammatory state. Of course, this is easier said than done. More-
over, positive cultures do not automatically conrm infection, and
great discretion must be exercised in determining how the micro-
biologic information obtained should be used. For example, tracheal
aspirates from intubated patients routinely reveal gram-negative ora,
but this nding in no way conrms pneumonia. A single blood cul-
ture growing S. epidermidis is similarly difcult to interpret.
Despite ubiquitous use in surgical patients, there is not a great
deal of evidence in the clinical literature to substantiate the effec-
tiveness of either empirical or streamlined antibiotic therapy.
Further efforts must be made to nd such evidence because this
practice could theoretically exact a substantial cost from both the
patient and the environment in which the patient is cared for. In
the meantime, clinicians must approach the development of SIRS
or other nonspecic signs of infection in their patients by predi-
cating antimicrobial use on carefully formulated diagnostic crite-
ria for the presence of infection; CDC consensus denitions of
infection are a good starting point.
Source Control for Management of SIRS
In the past, because the term sepsis was loosely used to describe
any general systemic inammatory state and because such states
often arose as a result of infection, it was assumed that antimicro-
bial therapy was generally appropriate in the management of sep-
tic patients. Now that sepsis has been redened exclusively as
SIRS in a patient in whom a causative source of infection has been
identied, it is clear that the use of antibiotics in septic (i.e.,
SIRS) patients should be more discriminating.
In the same vein, so-called source control was developed as a
strategy for managing septic patients. Like the term sepsis, the
term source control traditionally connotes management of infec-
tion rather than, more generally, management of a cause of inam-
mation. Classically, source control consists of a three-pronged
approach employing measures to (1) eradicate a focus of infection,
(2) eliminate ongoing microbial contamination, and (3) render the
local environment inhospitable to microbial growth and tissue
invasion. Diligent source control has long been considered pivotal
to successful management of sepsis. Although this traditional
approach addresses the infectious causes of local or systemic
inammation very well in a great variety of clinical situations, good
judgment must not yield to dogmatism in deciding which process-
es are required to control infection and inammation. Challenging
convention, a surprising number of investigators have successfully
managed many supposedly surgical conditions (e.g., appendicitis
22
and intra-abdominal abscess
23
) without intervention or by using
only prolonged antibiotic therapy. Seasoned general surgeons are
well aware that if acute cholecystitis is not operated on urgently, it
may certainly harm the patient or cause recurring discomfort, but
it may also resolve completely on its own. What actually consti-
tutes adequate source control and how this can be measured are
critical questions and remain the subject of debate. These ques-
tions become particularly problematic with respect to ICU
patients, in whom SIRS is highly nonspecic.
It would seem rational to take our current understanding of SIRS
as encompassing both infectious and noninfectious pathology and
extrapolate it to the concept of source control. Indeed, as regards
more complex surgical patients, it may be appropriate to broaden
the denition of source control to include control of all causes of
SIRS, not merely infectious ones. For example, debriding devitalized
injured tissue, removing a rejected allograft, and resolving postoper-
ative atelectasis are all important for successfully abating a systemic
inammatory state that might easily be mistaken for a manifestation
of infection. Deemphasizing infection as the predominant cause of
SIRS and withholding antibiotic therapy until stricter, more focused
diagnostic criteria for infection are met should make treatment para-
digms for managing difcult surgical patients, if not altogether more
effective, at least more evaluable.
If one assumes that source control is in fact a therapeutic
response to the presence of SIRS, one may then think of it as being
either assisted or unassisted. The therapeutic response is initiated
by the host, with either complete or partial success. Only in the lat-
ter instance should one assist the hosts efforts at source control by
providing antibiotics or taking surgical measures. This is by no
means to suggest that one should not search diligently for a cor-
rectable cause of infection or inammation but rather to suggest
that when such an effort yields negligible results, one should con-
sider the possibility that SIRS may be not only appropriate but
desirable and may represent the patients own adequate manage-
ment of the underlying physiologic insult. Thus, in certain situa-
tions, source control may be best regarded as an endogenous or
unassisted event. For instance, when fever, tachycardia, and leuko-
cytosis are observed in a surgical patient who is coping well, antibi-
otics should not necessarily be given automatically. Indeed, such
default therapy should be actively discouraged. One should also
keep in mind that some forms of injury or insult (e.g., some viral
infections) not only are very well managed without intervention but
may not even prompt clinically evident SIRS.
The notion that no intervention may be required is under-
standably difcult for many surgeons to embrace at the bedside.
Nonetheless, extensive ongoing research elucidating the complex
dynamic of circulating proinammatory and counterinammato-
ry mediators (e.g., TNF, the interleukins, and a host of other
cytokines) suggests that a poorly understood but highly sophisti-
cated biologic apparatus exists for responding to injury and insult.
Indeed, it is widely hypothesized (though yet unproved) that this
systemic response can be manipulated to restore health in stressed
or deteriorating surgical patients. The prospect of untangling the
complex biology of systemic inammation through advances in
this eld is truly engaging.
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 10
References
1. Marshall J, Sweeny D: Microbial infection and the
septic response in critical surgical illness. Arch
Surg 125:17, 1990
2. Meakins JL, Marshall JC: The gut as the motor of
multiple system organ failure. Splanchnic
Ischemia and Multiple Organ Failure. Marston A,
Ed. Edward Arnold, London, 1989, p 339
3. Goris RJA, Boekhorst TAP, Nuytinck JKS, et al:
Multiple organ failure: generalized autodestruc-
tive inammation? Arch Surg 120:1109, 1985
4. Bone RC, Fisher CJ Jr, Clemmer TP, et al: Sepsis
syndrome: a valid clinical entity. Crit Care Med
17: 389, 1989
5. Bone RC: Lets agree on terminology: denitions
of sepsis. Crit Care Med 19:973, 1991
6. Sibbald WJ, Marshall J, Christou N, et al: Sep-
sisclarity of existing terminology . . . or more
confusion? Crit Care Med 19:996, 1991
7. American College of Chest Physicians/Society of
Critical Care Medicine Consensus Conference:
Denitions for sepsis and organ failure and guide-
lines for the use of innovative therapies in sepsis.
Crit Care Med 20:864, 1992
8. Vincent J, Bihari DJ, Suter PM, et al: The preva-
lence of nosocomial infection in intensive care
units in Europe: results of the European
Prevalence of Infection in Intensive Care (EPIC)
Study. JAMA 274:639, 1995
9. Casadevall A: Crisis in infectious diseases: time
for a new paradigm? Clin Infect Dis 32:790,
1996
10. Lennard ES, Mineshew BH, Dellinger EP, et al:
Leukocytosis at termination of antibiotherapy:
its importance for intra-abdominal sepsis. Arch
Surg 115:918, 1980
11. Stone HH, Bourneuf AA, Stinson LD: Re-
liability of criteria for predicting recurrent sepsis.
Arch Surg 120:17, 1985
12. Adnet F, Borron SW, Vicault E, et al: Value of C-
reactive protein in the detection of bacterial con-
tamination at the time of presentation in drug-
induced aspiration pneumonia. Chest 112:466,
1997
13. Young B, Gleeson M, Cripps AW: C-reactive
protein: a critical review. Pathology 23:118, 1991
14. Cursons RTM, Jeyerajah E, Sleigh JW: The use
of polymerase chain reaction to detect sep-
ticemia in critically ill patients. Crit Care Med
27:937, 1999
15. The Society for Hospital Epidemiology of
America, the Association for Practitioners in
Infection Control, the Centers for Disease
Control, the Surgical Infection Society:
Consensus paper on the surveillance of surgical
wound infections. Infect Control Hosp Epide-
miol 13:599, 1992
16. Stamm WE, Hooton TM: Management of uri-
nary tract infection in adults. N Engl J Med
329:1328, 1993
17. Stark RP, Maki DG: Bacteriuria in the catheter-
ized patient: what quantitative level of bacteri-
uria is relevant? N Engl J Med 311:560, 1984
18. Bullard KM, Dunn DL: Diagnosis and treat-
ment of bacteremia and intravascular catheter
infections. Am J Surg 172(suppl 6A):13S, 1996
19. American Thoracic Society: Hospital-acquired
pneumonia in adults: diagnosis, assessment of
severity, initial antimicrobial therapy, and pre-
ventive strategies: a consensus statement. Am J
Respir Crit Care Med 153:1711, 1995
20. Sterling TR, Ho EJ, Brehm WT, et al: Diag-
nosis and treatment of ventilator-associated pneu-
monia-impact on survival. Chest 110:1025,
1996
21. Timsit M, Misset B, Renaud B, et al: Effect of
previous antimicrobial therapy on the accuracy
of the main procedures used to diagnose noso-
comial pneumonia in patients who are using
mechanical ventilation. Chest 108:1036, 1997
22. Eriksson S, Granstrom L: Randomized control-
led trial of appendectomy versus antibiotic ther-
apy for acute appendicitis. Br J Surg 82:166,
1995
23. Montgomery RS, Wilson SE: Intraabdominal
abscess: image-guided diagnosis and therapy.
Clin Infect Dis 23:28, 1996
2003 WebMD Inc. All rights reserved.
8 Critical Care
ACS Surgery: Principles and Practice
14 DIAGNOSIS OF INFECTION 11