Clinical review of the management of fulminant clostridium difficile infection.

Abstract

Clostridium difficile infection (CDI) is a frequent cause of morbidity and mortality among elderly hospitalized patients. A small but increasing number of patients have developed fulminant CDI, and a significant number of these patients require emergency colectomy. In this review, we discuss the risk factors, pathophysiology, diagnosis, and management of fulminant CDI.
A literature search (Medline, Embase, Cochrane Library, Biosis, Science Citation Index, Ovid Journals) was performed from the period between January 1980 and June 2008 using the key words "Clostridium difficile,"pseudomembranous enterocolitis,"colectomy,"acute abdomen,"antibiotic-associated diarrhea," or "fulminant Clostridium difficile colitis." Articles not in English or not related to human subjects were excluded. For this review, we analyzed the articles identified in our original search and those articles cited in the original review articles. No randomized trials were found on the surgical management of fulminant CDI and only retrospective studies with a minimum of five patients were used in the review. With respect to medical treatment, we based our review on guideline articles, systematic reviews, and available randomized trials.
Both the incidence and severity of CDI are increasing. Fulminant CDI is underappreciated as a life-threatening disease because of a lack of awareness of its severity and its nonspecific clinical syndrome. Early diagnosis and treatment are essential for a good outcome, and early surgical intervention should be used in patients who are unresponsive to medical therapy. The surgical procedure of choice is a total abdominal colectomy with end ileostomy, although the mortality rate remains high.

Full text

Page 1

American Journal of Gastroenterology
C ?2008 by Am. Coll. of Gastroenterology
Published by Blackwell Publishing
ISSN 0002-9270
doi: 10.1111/j.1572-0241.2008.02198.x
Clinical Review of the Management of Fulminant
Clostridium difficile Infection
M. Raffat Jaber, M.D.,1Snorri Olafsson, M.D., Ph.D., M.P.H.,2Wesley L. Fung, M.D.,3and
Mark E. Reeves, M.D., Ph.D.3
1Department of Surgery and Department of Medicine;2Division of Gastroenterology and Department of
Medicine; and3Department of Surgery, Loma Linda University Medical Center, Loma Linda, California
CME
BACKGROUND:Clostridium difficile infection (CDI) is a frequent cause of morbidity and mortality among elderly
hospitalized patients. A small but increasing number of patients have developed fulminant CDI, and
a significant number of these patients require emergency colectomy. In this review, we discuss the
risk factors, pathophysiology, diagnosis, and management of fulminant CDI.
A literature search (Medline, Embase, Cochrane Library, Biosis, Science Citation Index, Ovid
Journals) was performed from the period between January 1980 and June 2008 using the key
words “Clostridium difficile,” “pseudomembranous enterocolitis,” “colectomy,” “acute abdomen,”
“antibiotic-associated diarrhea,” or “fulminant Clostridium difficile colitis.” Articles not in English or
not related to human subjects were excluded. For this review, we analyzed the articles identified in
our original search and those articles cited in the original review articles. No randomized trials were
found on the surgical management of fulminant CDI and only retrospective studies with a
minimum of five patients were used in the review. With respect to medical treatment, we based our
review on guideline articles, systematic reviews, and available randomized trials.
Both the incidence and severity of CDI are increasing. Fulminant CDI is underappreciated as a
life-threatening disease because of a lack of awareness of its severity and its nonspecific clinical
syndrome. Early diagnosis and treatment are essential for a good outcome, and early surgical
intervention should be used in patients who are unresponsive to medical therapy. The surgical
procedure of choice is a total abdominal colectomy with end ileostomy, although the mortality rate
remains high.
(Am J Gastroenterol 2008;103:3195–3203)
DATASOURCES:
CONCLUSION:
INTRODUCTION
Clostridium difficile infection (CDI) is one of the most com-
mon nosocomial infections and a frequent cause of morbid-
ity and mortality among elderly hospitalized patients (1).
Approximately 3 to 8% of CDI patients develop fulminant
disease, defined as patients whose course is complicated by
perforation, severe ileus with toxic megacolon, hypotension
requiring pressors, or refractory septicemia. A significant
number of these patients require emergency colectomy (2).
In this review, we discuss the risk factors, pathophysiology,
diagnosis, and management of fulminant CDI.
EPIDEMIOLOGY
CDI is the most common cause of nosocomial infectious di-
arrhea in adults (3). Both the frequency and severity of CDI
areincreasing(4–9).Intheearly2000s,reportsfromQuebec,
Canada, noted an increase in the frequency and severity of
CDI (6, 7, 10–13). Pepin et al. (6) reported that the incidence
ofCDImorethanquadrupledfrom35.6to156.3per100,000
To access a continuing medical education exam for this article, please visit
www.acg.gi.org/journalcme.
between 1991 and 2003, and the proportion of patients who
died within 30 days after diagnosis increased from 4.7% in
1991 to 13.8% in 2003. The incidence of CDI in patients
aged 65 yr or older increased tenfold between 1991 and 2003
(6, 11). The disease seems to be more serious and refrac-
tory to therapy than it was in the past according to multiple
reports, as indicated by increased rates of toxic megacolon,
need for colectomy, associated shock, or death (6, 10–12).
The increase in mortality attributed to CDI was 16.7% be-
tween 2003 and 2004 alone (6, 11). Among those admitted
to the hospital in the region, the incidence increased from 3
to 12 per 1,000 persons between 1991 and 2002 and from
25 to 43 per 1,000 persons between 2003 and 2004 (6). A
recent retrospective observational cohort study from Quebec
(12) reported 165 cases of CDI that necessitated admission
to the intensive care unit (ICU) or prolongation of ICU stay
overa30-monthperiod(January2003toJune2005).Ofthese
165 patients, 38 patients required an emergency colectomy.
Generally, the disease was more frequent, more severe, more
refractorytotherapy,andsubjecttohighratesofrelapse(6,7,
9–13)inthatregion.Fivevariableswereassociatedwiththese
complications: age older than 65 yr, leukocyte count higher
than 20 × 109cells/L, acquisition of infection in a hospital,
3195

Page 2

3196 Jaber et al.
renal failure, and immunosuppression (6). In the United
States, the Center for Disease Control and Prevention re-
ported an increasing frequency and severity of CDI based on
reports from U.S. physicians. In addition, a review by Mc-
Donald et al. (9) of the International Classification of Dis-
eases, Ninth Revision (ICD-9) coding revealed an increase
in the United States from 82,000 cases in 1996 to 178,000 in
2003. Dallal et al. (5) reported that the incidence of CDI in
hospitalized patients almost doubled from 0.68% in 1990 to
1.2% in the year 2000. The development of life-threatening
symptomsalsodoubledfrom1.6%in1990to3.2%in2000in
the same group. In Pittsburgh, 1.1% (27 of 2,540) of patients
diagnosed with nosocomial CDI from 1989 to 1999 needed a
colectomy(5)whilethisincreasedtoaremarkable10.3%(26
of 253) of patients from 2000 to 2001 (some of whom had
been referred from other facilities because of severe CDI)
(14). Dial et al. (15) reported that 19% of 94 patients with
CDI required ICU admission, 9% required colectomy, 13%
developedrenalfailure,22%relapsed,and22%died.In2002,
Morrisetal.(16)reportedthattheoverallmortalityratefrom
CDI was 15.3%, a significant increase from the 3.5% mor-
tality rate in their previous study in 1995 (17).
BI/NAP1 (also known as NAP1/027) is a highly toxigenic
strain that was historically uncommon. This strain produces
15–20 times more toxin than more common strains. In addi-
tion,BI/NAPIisnowthoughttoberesponsibleforseverehos-
pitaloutbreaksandisbecomingmorefrequentincommunity-
associated disease. It has caused nosocomial and community
outbreaks in North America, Great Britain, and the Nether-
lands (9–12, 18, 19). Together, these data indicate that CDI
is evolving into a much more serious disease with significant
recent increases in severity, morbidity, and mortality.
Table 1. Summary of Special Characteristics and Outcomes in Patients Who Underwent Surgery for Fulminant CDI
CT Scan
Type Total
(n)
Median
WBC × 109/L Surgery (%)
28
–‡
32§
28
23
27.2
36.1
34.02
35
Recent Undiagnosed Prior
to Surgery (%)
No
Mortality
(%)∗
Study, Reference of Study Year
Diarrhea (%) (n)†Positive (%)
Hermsen (95)
Hall (96)
Ali (97)
Byrn (62)
Koss (46)
Longo (4)
Dallal (5)
Synnott (54)
Grundfest-
Broniatowski
(41)
Jobe (17)
Trudel (57)
Lipsett (43)
Medich (53)
R
R
R
R
R
R
R
R
2008
2008
2008
2008
2006
2004
2002
1998
1996
13
36
36
73
14
67
44
5
12
62
55
22
36
57
45
75
40
–
–
–
–
–
–
33
16
21
37
20
–
25
8 62
94
97
46
36
47
34
36
48
57
80
42
32
32
35
9
46
39
–
–
14
50
13
11
–
33
100
89
98
100
–
– R, RV
R
R
R
R
1995
1995
1994
1992
10–
–
–
–
30––
–
11
– 30
71
38
33
70
–
90
0
8
–∗∗
100
100
13
10
27.3
–††
62
40506
R = retrospective; RV = review; WBC = white cell count; CT = CT scan.
∗Post operative 30-day mortality.
†Total number of patients who underwent a CT scan of the abdomen.
‡WBC = 33 ± 19 th/cm2(survivors) and 37 ± 20 th/cm2(nonsurvivors) (P < 0.61).
§WBC = 23.3 × 109/L (survivors) and 40.8 × 109/L (nonsurvivors) (P < 0.01).
∗∗X-rays showed transverse colon dilation and loss of haustrations, 72% of patients in this series, with a mean diameter of 9.9 ± 3.4 cm.
††50% of patient had WBC > 30.0 × 109/L.
RISK FACTORS
The most important risk factor for development of CDI is
prior antibiotic use up to 8 wk previously (16, 20, 21).
Many antibiotics have been implicated (20–23), but those
with the highest risk are clindamycin (24), cephalosporins
(especially ceftriaxone) (25, 26), and fluoroquinolones (12)
(ciprofloxacin(27,28),levofloxacin(14,29,30),gatifloxacin
(14)).Fluoroquinoloneusehasbeenidentifiedasanindepen-
dent predictor of a positive C. difficile toxin assay in hospi-
talized patients (30, 31). Muto et al. (14) reported a large
outbreak of CDI with an unexpected proportion of colec-
tomies and deaths following increased fluoroquinolone use.
CDIoccurscommonlyinpatientswithsevereunderlyingdis-
ease and in hospitalized adults who receive antibiotic ther-
apy (15–25%), including those who receive even one dose of
prophylaxis for surgery (25, 32–37). Kreisel et al. reported
that 6% of a total of 357 patients with positive C. difficile
toxin received prophylactic antibiotic therapy only for elec-
tive surgical procedures (35). Jobe et al. (17) reported that
preoperative antibiotic use was the most frequent predispos-
ing factor for CDI, accounting for 25% of a total of 201
patients with CDI. In a study of 67 patients who had a colec-
tomy for fulminant CDI at various Veterans Administration
hospitals, Longo et al. (4) reported that 54% developed it
during a hospitalization for an unrelated illness, and 87% af-
ter a surgical procedure. Similarly, Dallal et al. (5) reported
that 75% of patients who required a colectomy for fulminant
CDI had a recent surgical procedure or were on a surgical
service (Table 1). Other risk factors associated with CDI in-
clude uremia, burns, chronic obstructive pulmonary disease,
cancer, abdominal surgery, cesarean section, antiperistaltic
medications, proton pump inhibitors, stay in an intensive

Page 3

Management of Fulminant Clostridium difficile Infection3197
care unit, longer hospital stay, chemotherapy, and postpy-
loric tube feedings (2, 4, 15, 25, 26, 38, 39). Renal disease
has been associated with increased mortality in patients with
CDI(1,2,40).Multiplestudieshavedescribedanassociation
between an immunocompromised state and susceptibility to
thisopportunisticorganism,aswellaspooreroutcomesinthis
group (1, 2, 40–43). Immunosuppressed patients with solid
organ transplantation and those with an impaired antibody-
mediated immune response to C. difficile toxins as well as
patients with inflammatory bowel disease exacerbation are at
an increased risk of fulminant CDI (5, 44, 45, 46, 62). Pa-
tients with immunosuppression, prior history of successfully
treated CDI, and those who have undergone recent surgi-
cal procedures have the highest risk of developing fulminant
CDI (5, 46).
PATHOPHYSIOLOGY
C. difficile colonizes the human intestinal tract after the nor-
mal gut flora has been altered by antibiotic therapy. It is not
known why only a fraction of patients exposed to a given
antibiotic becomes colonized with C. difficile or why only a
fraction (approximately one-third) of those colonized devel-
ops symptoms. One study suggests that serum IgG antibody
to toxin A may be an important component (45). Enterotoxin
A and cytotoxin B, produced by C. difficile, induce an in-
flammatory process. Both toxins adhere to receptors on the
humancolonocytebrushborderandcausenecrosisandshed-
dingofthesecellsintothelumen.TheC.difficiletoxinstarget
Rho family proteins, a group of low molecular weight GTP-
bindingproteinsthatregulateactinfilamentsinallcells.Once
Rho proteins are inactivated by C. difficile toxins, actin fila-
mentsdisintegrateandthedamagedcellcannotfunction.This
condition ultimately leads to increased capillary permeabil-
ity and peristalsis (47–49). The inflammation may result in
toxic megacolon and perforation in fulminant cases (4, 50).
Systemic symptoms are often caused by toxin-induced in-
flammatorymediatorsreleasedintothecolon(tumornecrosis
factor α, interleukin-8, macrophage inflammatory protein-2,
and substance P) rather than bacteremia, colonic perforation,
or ischemia (12, 47, 51, 52). Lamontagne et al. (12) found
4.3% gram-negative bacteremia associated with CDI in 138
patients with CDI that required an admission to the ICU or
prolongation of stay there. The basis for the wide spectrum
of disease manifestations and the reason why some patients
suffer more severe symptoms than others are unclear. It is
believed that the ability to generate an antibody-mediated
response to clostridial toxins varies among patients.
CLINICAL MANIFESTATIONS AND SEVERITY ASSESSMENT
C. difficile infection can lead to manifestations as varied
as the asymptomatic carrier state to fulminant disease with
toxic megacolon and perforation (5, 6). Although diarrhea is
the hallmark of CDI, it may be absent secondary to severe
colonic dysmotility (Table 1) (4, 5, 17, 41, 44–46, 53), mak-
ingfulminantcolitisdifficulttodiagnose.FulminantCDIisa
systemic inflammatory syndrome that may include abdomi-
nalpainwithorwithoutdiarrhea,fever,hypotension,tachyp-
nea,andleukocytosis(4,5,17,26,41,43,46,50,53,54,62).
Abdominal signs range from distention to generalized ten-
derness with guarding and acute surgical abdomen (2, 5, 43,
46, 50, 55–58). Patients usually present with strikingly high
whitebloodcell(WBC)counts,oftengreaterthan20×109/L
(Table 1) (2, 4–6, 43, 46, 53, 54). CDI can account for sev-
eral serious complications, including perforation, prolonged
ileus, megacolon, and death (2). Life-threatening systemic
toxicity can develop in some patients despite appropriate and
timely medical therapy (59, 60).
Hemodynamic, respiratory, and urinary output data have
been used in the severity assessment of patients with CDI.
Dallal et al. (5) classified patients with systolic blood pres-
sure >100 mmHg, heart rate >90, moderate tachypnea, and
oliguria responding to volume as patients with moderate dis-
easeandthosewhorequirevasopressors,whohaveheartrate
>120, are in need of mechanical ventilation or have severe
oliguria as patients with fulminant disease. Important factors
forpredictingworseoutcomesinsuchseverelyillpatientsin-
clude: higher WBC count (4–6, 61), renal insufficiency (6),
the need of vasopressors preoperatively (5, 46), and age (5,
12). A lower survival rate in patients with fulminant CDI
was associated with requirement for preoperative vasopres-
sors(5,12,46),multiorganfailure(46),andhigherAPACHE
II and III scores (4, 46). In Dallal et al.’s study, age signifi-
cantly predicted survival (69 vs 60 yr) (P = 0.015) (5). No
patientyoungerthan50diedandnoneolderthan80survived
colectomy. In addition, APACHE III data were available for
20/46 patients with fulminant colitis. The median APACHE
III score in these patients was 114. Nonsurvivors (n = 15)
had a median score of 117 versus 88 for survivors. The dif-
ference was not statistically significant (P = 0.15). However,
APACHE III scores increased daily until colectomy was per-
formed, each day significantly different (P < 0.01). Pepin
et al. (6) reported that a high WBC count (20 × 109/L or
greater) and an elevated creatinine level were strongly as-
sociated with adverse outcomes [adjusted odds ratio (AOR)
4.8; 95% confidence interval (CI) (2.8–8.4) and 3.1; (1.8–
5.2) respectively.] Lamontagne et al. (12) reported several
factors associated with increased 30-day mortality, includ-
ing WBC count >50,000 × 109/L (AOR, 18.6; 95% CI,
3.7–94.7), serum lactate > 5 mmol/L (AOR, 12.4; 95% CI,
2.4–63.7), age >75 yr (AOR, 6.5; 95% CI, 1.7–24.3), im-
munosuppression (AOR, 7.9; 95% CI, 2.3–27.2), and shock
requiring vasopressors (AOR, 3.4; 95% CI, 1.3– 8.7). Koss
et al. (46) reported that 66% of patients with organ failure
died versus 33.3% in the no organ failure group. In a recent
study of 73 patients who underwent surgery and acquired
C. difficile colitis, preoperative vasopressor requirement (P
= 0.04; odds ratio, 5.0), mental status changes (P = 0.002;
oddsratio,12.6),andtreatmentlength(P=0.002;oddsratio,
1.4) remained significant predictors of mortality (62).

Page 4

3198Jaber et al.
DIAGNOSIS
Generally the diagnosis of C. difficile infection is established
byacombinationofclinicalandlaboratorydata.Occasionally
lower endoscopy is necessary. Unexplained leukocytosis in a
hospitalized patient should prompt a work-up for C. difficile
infection (61). The ‘gold standard’ for the diagnosis of CDI
isthestoolcytotoxinassaywithasensitivityof67–100%and
specificity of 85–100%. However, it may take as long as 1–3
days to get the test results (23, 63). This test has been largely
replaced by enzyme-linked immunosorbent assay (ELISA)
that detects toxin A or toxin B. The ELISA tests are less
expensive and faster (2–6 h) than the cytotoxicity assay, with
asensitivityof75–85%(8).ELISA(TOXA/BTEST)detects
the presence of both toxins (A and B) and has a sensitivity of
92.2–94.5% and a specificity of 100% when compared with
thecytotoxicityassay(64,65).TheuseofanELISAtestmay
reduce the time from onset of symptoms to diagnosis and
thecommencementoftreatment(66).Somecliniciansrepeat
the ELISA test on two to three occasions to decrease the
chances of false negative results. A recent study by Mohan
et al. revealed that less than 1% of diagnoses are changed by
repeating the test and thus are not warranted and are not cost
effective (67). However, the ELISA tests can be negative in
patients with fulminant disease (4, 5, 46, 53, 54). Dallal et al.
(5) reported that 12.5% of toxin assays in 64 patients with
fulminant CDI were falsely negative. Similarly, Longo et al.
(4) reported that 18% of 67 patients who required colectomy
for fulminant CDI had a negative C. difficile stool assay.
Highclinicalsuspicionisnecessaryinpatientswhopresent
withanunexplainedacuteabdomen,especiallyinelderlycrit-
ically ill patients and those who have had recent antibiotic
exposure (5, 17, 41, 53, 55, 56, 58). A strikingly high WBC
count with bandemia should increase the index of suspicion
(Table 1). It is noteworthy that although CDI frequently de-
velops during hospitalization and often occurs after surgical
procedures, it can also begin outside the hospital (3, 5). In
patients with fulminant CDI, a computed tomography (CT)
Figure 1. Abdominal computed tomography scans from patients with fulminant Clostridium difficile colitis reveal pancolitis thickening and
dilatation of the colon.
scanoftheabdomenwithorwithoutcontrast,togetherwithan
appropriate clinical scenario, can be very suggestive of CDI
although detection of C. difficile toxins in the stool is still
necessary for the definitive diagnosis (Table 1). Longo et al.
(4) found that a CT scan of the abdomen is the most sensitive
indicator of whether antibiotic-associated diarrhea is accom-
panied by severe colonic inflammation. In the Dallal study
(5) where 39 patients underwent CT scans, it was diagnostic
inall39caseswhencombinedwiththeclinicalscenario.Ab-
dominal CT scan findings usually include colonic wall thick-
ening, colonic dilation, and ascites (Fig. 1) (4, 5, 43, 46, 53,
54, 68–72). The sensitivity and specificity of abdominal CT
imaging in detecting colonic abnormalities in patients with
CDI has a range of 52–85% and 48–93%, respectively (68,
73–75). However, CT scan findings correlate poorly with ac-
tual severity (43, 72, 74). Proctosigmoidoscopy may identify
pseudomembranes in as many as 23% of mild cases and 87%
of fulminant cases (50, 76, 77). In the study by Longo et al.
(4), 26/67 of patients with fulminant CDI underwent sigmoi-
doscopy and it was positive in all. Dallal et al. (5) reported
that 10% of endoscopies in patients who died or underwent
colectomysecondarytofulminantCDIwerefalselynegative.
In addition, if the disease is localized to the right colon, sig-
moidoscopy will not be diagnostic and bowel prep may be
necessary to perform an adequate colonoscopy. There is a
risk of perforation with any colonoscopy or sigmoidoscopy,
but especially among the elderly, frail, debilitated population
of patients with severe colitis. Therefore, a CT scan may be
a safer test and should be considered first when severe CDI
is suspected.
TREATMENT (Fig. 2)
When the diagnosis of CDI is likely or confirmed, the impli-
cated antimicrobial agent or agents should be discontinued
immediately or changed if possible. Supportive care should
be instituted including fluid resuscitation and ventilatory

Page 5

Management of Fulminant Clostridium difficile Infection 3199
Figure 2. Treatment algorithms for CDI.
support if needed. In addition, physicians should avoid an-
tiperistaltics, including narcotics and loperamide. Patients
suffering from severe CDI (i.e., patients with toxic mega-
colon, leukemoid reactions (leukocyte count > 20 × 109
cells/L), renal failure, hypotension, septic shock, significant
abdominalpain,distentionwithileus)requiretreatmentinan
intensive care unit and empiric-specific antimicrobial ther-
apy should be initiated. Oral vancomycin is the preferred