Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 04/22/15 from IP address 128.122.253.228. Copyright ARRS. For personal use only; all rights reserved
1163
Treatment of Critically Ill Patients with Sepsis of Unknown Cause: Value
H Michael J. Le& Sanjay Saini1 James A. Brink1’2
Peter F. Hahn1 Joseph
F. Simeone1
Mary
C. Morrison1’3
David Rattner Peter R. Mueller1
of Percutaneous
H
Because of the difficulty severe intercurrent illness of the gallbladder, in the
Cholecystostomy
a trial
intensive-care
laboratory,
and
in diagnosing acute cholecystitis by clinical and imaging methods
of percutaneous
unit
with
radiologic
search
high fevers,
despite
antibiotic
count
patients,
vague
in
18
cholecystostomy
persistent,
therapy,
was
unexplained
showed
sepsis
no alternative
were
abdominal
in critically ill patients with or percutaneous aspiration
present
tenderness
source
performed after
of infection.
in all patients, in
1 1, and
in 24 patients
a complete
clinical, Persistent
with elevated
septic
shock
WBC
requiring
vasopressors in 15. Sonographically, all patients had distended, spherical gallbladders, six had gallstones, eight had wall thickening, three had pericholecystic fluid, and four had Murphy’s sign. All patients were seen by a senior abdominal surgeon, who agreed to
a trial
of
percutaneous
cholecystostomy.
Fourteen
patients
(58%)
responded
to
percutaneous cholecystostomy, as evidenced by a decrease in WBC count, defervescence, and the ability to be weaned off vasopressors. Bile cultures were positive in four patients.
Ten
patients
(42%)
did
not
respond
to
percutaneous
cholecystostomy;
five
eventually died of unrelated causes. A respiratory source of infection was eventually found in three of these 10 patients, with no proved source of infection in the remainder. No complications related to catheter insertion occurred in this group of patients. Bile leaks occurred in two patients when the percutaneous cholecystostomy catheter was removed, but without serious consequence. Our experience suggests that a lower threshold for performing percutaneous cholecystostomy in this difficult clinical subset of patients is worthwhile. AJR
156:1163-1166,
June
1991
Since its introduction, percutaneous cholecystostomy has become widely established as a safe and effective method of decompressing the gallbladder in patients with acute cholecystitis who are at high risk for surgery [1 -3]. Our early experience with 80 percutaneous cholecystostomies confirmed the safety, ease of execution, and low complication rate of this technique. Previous authors [1 2] have performed cholecystostomy in these high-risk patients, only when clinical, laboratory, or imaging studies or diagnostic gallbladder aspiration have suggested acute cholecystitis. However, as our experience with percutaneous cholecystostomy grew, a subset of patients emerged in whom the diagnosis of acute cholecystitis was difficult and the indications for percutaneous cholecystostomy were unclear. These were usually patients in the intensive care unit (ICU) who had multiorgan disease, persisting unexplained sepsis, and who were referred for abdominal CT or sonography to exclude cholecystitis (acalculous or calculous) or other intraabdominal source of sepsis. Our experience with this subset of patients, in accord with the literature, suggested that sonography, clinical examination, and biliary scintigraphy were not completely reliable in excluding acute cholecystitis in this clinical situation [4-8]. Similarly, percutaneous aspiration of the gallbladder with Gram staining and culture of bile, although initially promising [9], recently has been proved inadequate, with a diagnostic sensitivity of less than 50% [10]. ,
Received October 29, 1990; accepted after revision December 27. 1990. 1 Department of Radiology, Massachusetts General Hospital, dress reprint 2 Present diology,
32 Fruit St., Boston, MA 02114. Adrequests to P.R. Mueller. address: Mallinckrodt Institute of Ra-
Washington
S. Kingshighway
University Blvd.,
Medical
St. Louis,
3 Present address: Radiology Broward Medical Center, 201 pano Beach, FL 33064.
School,
510
MO 63110. Department, North Sample Rd., Pam-
Department of Surgery, Massachusetts eral Hospital, 32 Fruit St., Boston, MA 02114. 0361 -803X/91/1 566-1163 © American Roentgen Ray Society
Gen-
Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 04/22/15 from IP address 128.122.253.228. Copyright ARRS. For personal use only; all rights reserved
1164
LEE
For these reasons, and because these patients are at increased risk for acute acalculous cholecystitis, we decided to lower our threshold for performing percutaneous cholecystostomy in this group of patients. Percutaneous cholecystostomy was performed in 24 patients in the ICU who had persisting, unexplained sepsis to exclude acute cholecystitis as a possible source of infection. Findings on mobile sonographic examinations done to exclude cholecystitis were abnormal in some patients, but frequently sonography revealed only a distended gallbladder containing sludge. That finding is indicative of bile stasis, which frequently occurs in patients who have prolonged critical illness, and which carries an increased risk of bile inspissation, obstruction of the cystic duct, and acute cholecystitis [1 1]. The results, selection criteria integrating clinical data and sonographic findings, and potential complications of this approach in this group of patients are discussed.
ET AL.
AJA:156,
For
the
neous ance
and Methods
cystostomy,
Twenty-four
patients
sepsis
undergoing
of unknown
tion criteria
included
distended,
tense
or
gallbladder
failure
(six),
(i 2),
pancreatitis
were
on respiratory
teral
nutrition.
24
had elevated
tenderness.
all of the
of the
abdominal
patients
were
congestive
complete
other
septic
clinical, sources.
of sepsis
included
of tips
of central
chest
studies
were
either
raphy
was
levels
occurred
on liver
not
and
a continuing
tostomy Mobile
gallbladder its
bare
[1
and
but
to the
cholecystic
fluid,
Although
considered
and
the
sonography
clinical if the
status
was
gallbladder
this
CT
cultures; scans;
and
radiologic
Biliary
scintig-
of liver
enzyme
an experienced cholecys-
to exclude
used was
defined
patients
a distended, by
tense
An abnor-
measurements
of a spherical sonographic
performed
acute
to select
20 of 24 patients).
of Murphy’s
appeared
out other
patient.
Other
period,
only
ruled
and
with
such as gallstones, presence
or after
had
of percutaneous
appearance
shape.
All cho-
sputum
on all patients
not
be-
or tensing
four had normal results
criterion
by the
was in
and
we consulted
in
localize
performed
disease.
the remaining
(seen
pear
shock,
to exclude
elevation
cholecystostomy
rather
chole-
ment
decrease features.
cholecystostomies by using
Seldinger
with the
placed
into
sent
was
dilated,
under
the
under
system compatible
in the
remaining
stains
and
with
sonographic
with
critically 1 7 patients.
in size
radiology
departand
gallbladder
control. at the
guid-
through catheter
We now bedside
prefer
[i 2]. in the
In all cases,
the was
by using
guidance ill patients
the was
sonographic
cholecystostomy
fluoroscopic
the
percuta-
fluoroscopy the
in
into
seven
was inserted
cholecystostomies
trocar
Gram
and
first
needle
gallbladder
7- to 12-French
and
the
described
leakage
in the
patients,
sheath
gallbladder
is more for
from
performed
these
the
bile
experience,
of sonography
For
percutaneous
performed
ranged in our
into
previously
of
(0.097 mm) guidewire
tract the
to enter
made
possibility
used
an i 8-gauge
A 0.038-in.
sheath,
was
were
technique.
punctured
using
Early
a combination
transhepatically
guidance, the
Catheters
locking
placed
This
ICU
and
specimens
culture.
Results
pain or
hr before
tests
stable
performed
sludge
usual
24
negative,
nonspecific
sonographic
gallbladder
or size,
similar
were
of each
gesting acute cholecystitis, also.
cultures
response
percutaneous
improvement
percutaneous
therapy.
were
An attempt
to
cavity.
had
neous
were
paren-
palpation.
abdominal
the
containing
clinical
urine,
to evaluate
distended
opposed
blood,
or showed
was
percuta-
disappear-
an overall
24 hr after
for vasopressor
sonographic
3].
,
area
abdominal
was
of percutaneous was
catheters;
to a trial
undergo
mally
line
In all cases,
main
on time
or
a grimace
in all patients
agreed
The
assess
laboratory
who
gallbladder volume
and
20 patients;
sonography
within
catheters
under
techniques
and a
temperature
for septic
in the
surgeon,
cholecystitis. to
changed
Mild,
tests.
need
Cholecystostomy
technique
renal
total
(body
but
elicited
multiple
normal
in
to
at the
performed
All
on palpation,
occurring
to
All patients
receiving
fever
cholecystostomy
performed.
function
abdominal
not
radiographs.
response
defervescence,
a decrease
a single-stick
complex
tenderness
as
in WBC count to normal, and the ability to be weaned from vasopressors within 3 days. A negative response occurred when the patient’s clinical course was unaltered by percutaneous cholecystostomy, with persisting fevers, high WBC counts,
Selec-
(seven),
(one).
vasopressors
was
Tests
sources
and
difficult
imaging,
cultures serial
high
unconscious,
Percutaneous
after
ICU
had
was
were
were
failure
trauma
in the
antibiotics
these
had
and 1 1 had vague abdominal
musculature
procedure.
multiple
receiving
pain
receiving
cardiac
and
WBC counts, patients
lecystostomy; the
All patients
patients
Abdominal
cause
cause
a positive
defined
including respiratory failure or acute (21 patients), recent major abdominal
(five),
>1 01 #{176}F [38.3#{176} C]), i 5 were 18
at sonography.
ventilators
All
prospectively.
of unknown
syndrome
surgery
cholecystostomy
studied
sepsis
problems,
distress
thoracic
were
persistent
medical and surgical respiratory
percutaneous
origin
study,
was
well-being
to perform to treat
of this
of abdominal
in clinical
ance.
Subjects
purposes
cholecystostomy
June 1991
of
shape criteria
sug-
wall thickening, sign,
on other
were patients
acute
cholecystitis
normal
on sonography.
as
pen-
evaluated with was
a not
A dramatic improvement in the clinical condition of 14 patients was observed within 24 hr. Signs of improvement included defervescence of fever (1 4 patients) and disappearance of abdominal pain and tenderness (nine). In addition, WBC counts returned to normal in 1 1 patients, and vasopressors were discontinued in nine. Sonographic features, before cholecystostomy, in those patients who responded included distended gallbladders (14 patients), pericholecystic fluid (two), gallstones (three), wall thickening (five), and presence of Murphy’s sign (four; Table 1). Six of the responders had only a distended gallbladder with or without sludge, and one other patient had wall thickening in addition. The remaining patients had a combination of sonographic signs. One of the patients in the group who responded to treatment died 3 weeks after percutaneous cholecystostomy of severe congestive cardiac failure. Cholecystectomy was performed in one patient in this group and methyl tert-butyl ether (MTBE) dissolution therapy was performed in another when their clinical conditions stabilized (both had gallstones). Cholecystostomy was, therefore, the definitive therapy in 1 3 of these 1 4 patients. No clinical response was observed in the other 1 0 patients. Sonographically, before cholecystostomy, gallbladder distension and sludge were the sole imaging features in five patients, gallstones were present in three, wall thickening in three, and pericholecystic fluid in one (Table 1). Five of these patients eventually died of a combination of sepsis and multiorgan failure. A respiratory source of infection was eventually found in three of the 1 0 nonresponders, but no definitive source of infection was ever determined in the remaining seven patients.
PERCUTANEOUS
AJA:i56,
June 1991
TABLE
1: Sonographic
Findings
Before
Percutaneous
Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 04/22/15 from IP address 128.122.253.228. Copyright ARRS. For personal use only; all rights reserved
Cholecystostomy Number of Responders
Finding
Number of Nonresponders
(%) Gallstones Sludge
3 11
(21) (79)
3
(30)
Wall thickening
5 (36)
Penicholic Murphy’s
2 4
1 0
Total
(14) (29)
14 (100)
(#{176}/)
(%) 9 (90) 3 (30)
fluid sign
Total Number
(10) (0)
10 (100)
6
(25)
20 (85) 8 (33) 3 (12) 4
(17)
24 (100)
One patient had a laparotomy for suspected ischemic bowel disease, at which the gallbladder appeared normal. No grossly purulent material was aspirated from the gallbladder in any patient in this series. Bile cultures were positive in four patients and blood cultures in one. The duration of percutaneous catheter drainage varied between 1 0 days to 3 months (mean, 1 .3 months). No complications related to catheter insertion occurred in this series. Bile leaks, related to catheter withdrawal, occurred in two patients. Abdominal pain subsided in one patient after 24 hr of observation. The catheter was immediately reinserted into the gallbladder in the second patient, when contrast material was seen leaking into the abdominal cavity during catheter removal under fluoroscopic control.
Discussion Detection of sepsis originating from the gallbladder (calcubus or acalculous cholecystitis) is particularly difficult in patients who are unresponsive and have multiorgan failure [48]. Clinical and biochemical markers are often nonspecific in the presentation of acute acalculous or calculous cholecystitis in this situation and contribute to the delay in diagnosis and treatment [1 1 ]. Pain and tenderness in the right upper quadrant can be absent. Moreover, it is often difficult to elicit clinical signs and symptoms in patients who are on respirators and have decreased mental awareness. Similarly, the accuracy of sonography in the diagnosis of acute acalculous cholecystitis has been reported to be as low as 58% [1 3]. Sonographic signs such as wall thickening and pericholecystic fluid, which were seen in some patients in this series, are often nonspecific indicators of acute cholecystitis and may be seen with various entities such as ascites, hypoalbuminemia, and hepatitis, among others [14]. Cholescintigraphy with 99mTciminodiacetic acid, although highly accurate in the diagnosis of acute cholecystitis (95%) under normal circumstances [1 5], has a high false-positive rate in patients who have fasted for longer than 1 4 hr (false-positive in all four patients studied) [6], have severe intercurrent illness (50% false-positive rate) [7], or are receiving total parenteral nutrition, as were all the patients in this series (92% false-positive rate) [8]. Initial reports suggested that percutaneous gallbladder aspiration and analysis of bile would be helpful in confirming the
1165
CHOLECYSTOSTOMY
gallbladder as the source of sepsis [9]. However, recent reports have questioned the efficacy of percutaneous bile aspiration in these critically ill patients, who often are already receiving antibiotics. McGahan and Lindfors [1 0], in a recent study, estimated the sensitivity of Gram stains or cultures on percutaneous bile aspirate as less than 50%. The presence of WBCs signifies infection; however, their absence is not reliable for excluding infection in patients receiving antibiotics. In addition, cultures are definitive when positive, but are not available immediately. Therefore, negative Gram stains or cultures are meaningless in this population of patients. Because of this lack of sensitivity in the diagnosis of acute cholecystitis in these critically ill patients, and the need for early diagnosis and treatment, we elected to perform a ther-
apeutic trial of direct percutaneous subset of patients. Selection criteria sis of unknown
origin
after
other
cholecystostomy in this included persistent sepsources
of sepsis
were
excluded and gallbladder distension, often with retained sludge, as defined by sonography. Acalculous cholecystitis is known to occur more frequently in these patients with prolonged critical illness, who have an increased risk of bile inspissation, obstruction of the cystic duct, and acute cholecystitis [10]. Our results emphasize the poor sensitivity of sonography in diagnosing
cholecystitis
in these
patients.
Gallstones
were
present equally in both responders and nonresponders, and one patient who did not respond had a distended gallbladder, sludge, gallstones, wall thickening, and pericholecystic fluid (Table 1). Murphy’s sign was the most specific sonographic sign in that all four patients in whom the sign was elicited responded to percutaneous cholecystostomy. However, the absence of Murphy’s sign on sonography did not preclude a positive
response
to percutaneous
cholecystostomy.
Clini-
cally, the presence of abdominal tenderness (1 1 of 24 patients) was strongly suggestive of cholecystitis in this study, with nine of 1 1 patients responding to percutaneous cholecystostomy. However, abdominal pain or tenderness may be hard to elicit in ventilated ICU patients, as in five of our responders who did not have any clinical signs or symptoms of cholecystitis. One would expect a high failure-of-response rate using this approach, and indeed, 10 (42%) of 24 patients in this series did not respond
to percutaneous
cholecystostomy.
Our con-
cern about this approach is that the nonspecificity of selection criteria (both imaging and clinical) used might lead to unnecessary
procedures
performed
on the gallbladder
in this subset
of patients. However, percutaneous cholecystostomy was performed only after both the clinical and the radiologic data were reviewed with an experienced abdominal surgeon and after a careful search had excluded other sources of sepsis. Thus, percutaneous cholecystostomy was not performed randomly on all patients in the ICU who might have only one abnormality, such as a distended gallbladder. Still, because of the nonspecific
selection
criteria,
placement
of catheters
in
patients with normal gallbladders was unavoidable. An equivalent dilemma in general surgery is the removal of a certain percentage of normal appendices in patients with abdominal pain to avoid the grave error of misdiagnosing patients with
1166
Downloaded from www.ajronline.org by NYU Langone Med Ctr-Sch of Med on 04/22/15 from IP address 128.122.253.228. Copyright ARRS. For personal use only; all rights reserved
acute
LEE
appendicitis.
cholecystostomy
However, did
have
in this series, a significant
percutaneous
positive
effect
to ensure
patency
of the
cystic
duct
a distended,
tense
gallbladder
1 . McGahan
2.
3. 4.
5.
and
common bile duct before removing the catheter. Injection of contrast media through the catheter during removal also helps to outline the percutaneous tract and ensure its maturity. In summary, we believe that a trial of percutaneous cholecystostomy in critically ill ICU patients with persisting, unexplained fever and a distended, tense gallbladder on sonography is worthwhile, because of the often nonspecific clinical and sonographic findings of acute cholecystitis in these patients. One must first exclude all other potential sources of sepsis before contemplating percutaneous cholecystostomy. Sonographically,
AJA:i56, June 1991
REFERENCES
for
these patients in that surgical cholecystostomy or cholecystectomy was avoided and the gallbladder was eliminated as the source of sepsis. Importantly, no serious cholecystostomy-related complications occurred in these patients. Bile leaks, without serious clinical consequence, did occur in two patients in this series when the catheters were removed after 10 days. To minimize this risk, we now leave catheters in situ for a minimum of 2-3 weeks and perform cholecystocholangiography
ET AL.
with Murphy’s
sign in the presence of abdominal pain suggests the presence of acute cholecystitis and strongly predicts a positive response to percutaneous cholecystostomy. A positive response to percutaneous cholecystostomy may also occur in patients with no localizing signs, because of the difficulty in eliciting clinical data in these patients. If no response occurs, surgery has been avoided and the gallbladder has been eliminated as the cause of sepsis. The low complication rate found in this and other series [1 -3, 16], coupled with the fact that percutaneous cholecystostomy can be performed quickly and safely at the bedside, makes it an ideal technique for critically ill patients in the ICU.
6.
7.
8.
9. i 0. 11. 1 2. 13.
14. 1 5.
i 6.
JP, Lindfors
KK. Percutaneous
cholecystostomy:
an alternative
to surgical cholecystostomy for acute cholecystitis. Radiology 1989; 173:481-485 Vogelzang AL, Nemcek AA. Percutaneous cholecystostomy: diagnostic and therapeutic efficacy. Radiology 1988:168:29-34 Lindemann SR, Tung G, Silverman SG, Mueller PA. Percutaneous cholecystostomy: a review. Semin Intervent Radio! 1988:5 : 1 79-i 85 Laing FC, Federle MP, Jeffrey RB, Brown 1W. tJtrasonic evaluation of patients with acute right upper quadrant pain. Radiology 1981;140: 449-455 Samuels BL, Freitas JE, Bree AL, Schwab RE, Holler JT. A comparison of radionuclide hepatobiliary imaging and real-time ultrasound for the detection of acute cholecystitis. Radiology 1983:147:207-210 Larsen MJ, Klingensmith WC Ill, Kuni CC. Radionuclide hepatobiliary imaging: nonvisualization of the gallbladder secondary to prolonged fasting. J Nuc! Med 1982:23: 1003-1 005 Kalff V, Froelich JW, Lloyd A, Thrall JH. Predictive value of an abnormal hepatobiliary scan in patients with severe intercurrent illness. Radiology 1983:146: 191 -1 94 Shuman WP, Gibbs A, Audd TG, Mack LA. PIPIDA scintigraphy for cholecystitis: false positives in alcoholism and total parenteral nutrition. AJR 1982;138:i-5 McGahan JP, Walter JP. Diagnostic percutaneous aspiration of the gallbladder. Radiology 1985:155:619-622 McGahan JP, Lindfors KK. Acute cholecystitis: diagnostic accuracy of percutaneous aspiration of the gallbladder. Radiology 1988:167:669-671 Long TN, Heimbach DM, Camco CJ. Acalculous cholecystitis in critically ill patients. Am J Surg 1987:136:31-36 McGahan JP. A new catheter design for percutaneous cholecystostomy. Radiology 1988:166:49-52 Mirvis SE, Vainright JA, Nelson AW, et al. The diagnosis of acute acalcubus cholecystitis: a comparison of sonography, scintigraphy, and CT. AJR 1986;i47:1171-i175 Cooperberg PL, Gibney AG. Imaging of the gallbladder 1987. Radiology 1987:163:605-613 Weissman HS, Frank MS, Bernstein LH, Freeman LM. Rapid and accurate diagnosis of acute cholecystitis with 99m-Tc-HIDA cholescintigraphy. AJR 1979:132:523-528 vanSonnenberg E, Wittich GA, Casola G, et al. Diagnostic and therapeutic percutaneous gallbladder procedures. Radiology 1986:160:23-26