Clostridium difficile-associated disease: a persistently plaguing problem
Dale N Gerding
Chicago VA Healthcare System and Northwestern University Medical School, Chicago, Illinois, USA
Clostridium difficile-associated disease (CDAD) is the quintessential
antibiotic-related illness. Originally described in the preantibiotic era, this infection is today virtually nonexistent
in the absence of antibiotic use. (1) Detailed knowledge of its epidemiology and pathogenesis has continued to
evolve since the identification in 1977 of C. difficile as the cause of antibiotic-associated colitis.(1) Unfortunately, despite
this increased knowledge, the incidence of CDAD appears to be increasing. It is clear that we still lack the critical
information regarding diagnosis, epidemiology, and pathogenesis needed to design more effective preventive and
At least three "hits" or conditions are considered necessary for CDAD to develop. Most investigators
agree that two components are essential: exposure to antimicrobials and exposure to toxigenic
C. difficile, in that order. Because the majority of patients
who meet these first two criteria do not become ill, at least one additional factor is necessary. The additional
factor may be related to host susceptibility or immunity, to the virulence of the particular
C. difficile strain, or to the type and timing of the antimicrobial
exposure. Molecular-typing studies clearly show that even the most virulent C.
difficile strain produces asymptomatic colonization more
often than CDAD, suggesting that something other than virulence is necessary for CDAD. (2,3)
The vast majority of cases of CDAD occur in hospitals and similar institutions, presumably because of the higher
incidence of antimicrobial use in these settings, as well as the existence of a contaminated environment and the
presence of more patients with CDAD or asymptomatic C. difficile colonization. (3,4) CDAD is the most frequently
identified cause of nosocomial diarrhea. Several studies have shown that to be cost-effective diagnostic laboratory
testing for diarrhea occurring in a hospital setting should initially be confined to tests for C. difficile.(5)
Exposure to any antimicrobial can result in CDAD, but certain agents are more likely to be associated with it,
e.g., clindamycin, ampicillin, and third-generation cephalosporins. Two publications (6,7) suggest that the risk
of CDAD is much higher with the use of third-generation cephalosporins (ceftriaxone, ceftazidime) than with ticarcillin/clavulanic
acid when used during the same time period in the same institutions. This observation is potentially helpful if
an institution is experiencing high rates of CDAD in association with the use of cephalosporins and desires to
change institutional antimicrobial policy.
Although CDAD is rarely diagnosed in the outpatient setting (7.7 cases per 100,000 person-years of observation),
it may be because diagnostic testing is not performed often enough in outpatient settings to detect CDAD.(8,9)
Riley and colleagues found the diagnosis of C. difficile in episodes of community - acquired diarrhea in Australia increased from
2.6 to 10.7% of submitted specimens after an educational program for general practitioners advised including C. difficile testing
for outpatients with diarrhea.(9) There is a need for similar studies of community-acquired diarrhea in the United
The diagnosis of CDAD is complicated by the lack of a single sensitive and specific test. Tests for cytotoxin (toxin
B) or toxin A in stool specimens are highly specific but less sensitive than a stool culture for C. difficile, which
is sensitive but lacks specificity because C. difficile may be colonizing but not causing CDAD. (5,10) Sensitivity of the toxin
assays can be improved by testing several stool specimens; specificity of the stool culture can be increased by
performing toxin assays in vitro on all C. difficile isolates if the stool cytotoxin assay is negative or was not performed.
Both alternatives are time consuming, and performing both toxin assays and C.
difficile culture on stool specimens is an added expense
and burden for already busy clinical laboratories. A single diagnostic test for CDAD that has both high sensitivity
and specificity is needed, preferably one that is rapid and inexpensive as well.
The clinical criteria for who should be tested for CDAD also need better definition. Empirical recommendations
for testing include a history of antimicrobial use within 2 months and that laboratories test only liquid or unformed
stools from patients suspected of having CDAD (ensuring that diarrhea is present).(5) A clinical prediction rule
for CDAD consisting of prior antibiotic use (within 30 days) and appreciable diarrhea (more than 3 partially formed
or watery stools in a 24-hour period) has been proposed and tested using the C.
difficile stool cytotoxin assay, but it has not been validated.(11)
In the initial derivation model, this rule was 80% sensitive and 45% specific, with a positive predictive value
of only 18%. The major value of the rule was its negative predictive value (94%), which would have precluded 39%
of the cytotoxin testing requested and missed the diagnosis in only 6% of patients whose stools were positive for
cytotoxin. If validated, considerable savings could be achieved by eliminating unnecessary testing through this
screening protocol. For patients whose stools might have been positive but were not tested, testing can be reconsidered
if symptoms persist.
Treatment and the Risk of VRE
The clinical efficacies of metronidazole (250 mg 4 times daily) and of vancomycin (125 mg 4 times daily), each
for a total of 10 days, are considered to be equivalent for the treatment of CDAD.(5) However, the increased incidence
of infections with vancomycin-resistant enterococci (VRE) in the 1990s has caused considerable concern regarding
the use of oral vancomycin in hospitals and the subsequent risk of VRE infection. A small number of studies have
implicated the use of oral vancomycin as a risk factor for VRE infection or colonization when compared to control
patients, but the number of patients administered oral vancomycin was less than 25% of the patients with VRE in
the studies. In contrast, use of intravenous vancomycin was found to be a significant risk in many more such studies
and in a much higher percentage of patients with VRE.(12) Four of 5 controlled studies of the risk of VRE infection
that looked at the use ofmetronidazole found it to be a significant risk factor; however, the indications for metronidazole
were not stated so it is not possible to determine to what extent metronidazole use was for the treatment of CDAD.
The 1995 recommendations of the Hospital Infection Control Practices Advisory Committee (HICPAC) for preventing
the spread of vancomycin resistance discourage the routine use of vancomycin as initial therapy for CDAD.(13) Because
of the lack of an effective therapy for severe VRE infections, it seems prudent to restrict all unnecessary use
of vancomycin. The potential for the use of oral metronidazole as a treatment of CDAD to also increase the risk
of VRE infection should be examined in future epidemiologic studies.
In addition to the vancomycin and metronidazole link between CDAD and VRE, these infections share a number of other
risk factors, including nosocomial origin, older patient age, severity of illness, length of hospital stay, use
of rectal thermometers, enteral feedings, treatment with antibiotics, number of antibiotics received, and exposure
to clindamycin, imipenem, and third-generation cephalosporins.(12) Both infections are associated with high rates
of asymptomatic gastrointestinal colonization, environmental contamination, and carriage of the organisms on the
hands of health care personnel.
Prevention and Control
Two strategies have been employed to control the rate of CDAD in hospitals: prevention of horizontal spread of
the C. difficile organism
using traditional infection control barrier methods and reduction of the risk of symptomatic disease (CDAD) after
exposure to C. difficile.(5)
Although the results of these efforts have generally been discouraging, several notable successes have been achieved,
including the demonstration of the benefit of personnel wearing gloves when handling body substances and the reduction
of risk when electronic rectal thermometers were replaced by disposable thermometers.(5)
Similarly, efforts to reduce the risk of CDAD after exposure to C.
difficile have also been somewhat successful, especially
restriction of the use of specific antimicrobial agents that have been identified as risk factors, such as clindamycin.(5)
Numerous prophylactic measures have been used during antimicrobial treatment, but only the yeast Saccharomyces boulardii
has shown much promise as a preventive agent. Prophylactic use of lactobacilli and yogurt have not been shown to
be effective in preventing CDAD in humans.
My personal view is that traditional infection control measures have been of limited success with CDAD for two
reasons: patients are targeted for barrier precautions too late (after CDAD has developed) and personnel fail to
rigorously follow infection control practices, especially with regard to handwashing and gloving. Future efforts
should focus on two areas for prevention of CDAD. First, everyone who prescribes antimicrobials must practice good
antimicrobial use stewardship. As a steward manages another's property or possessions, the good antimicrobial steward
considers carefully whether antimicrobials are needed, prescribes only the agent and dosage appropriate for the
infection being treated and is cognizant of the potential long-term adverse effect of imprudent antimicrobial use
on the preservation of future efficacy. Reducing the number of patients who needlessly receive antimicrobials will
unquestionably reduce the number of patients susceptible to CDAD.
The second area of focus should be on the use of prophylactic agents to prevent CDAD if a patient is exposed to
C. difficile after
receiving antimicrobials. Biologic agents, such as S. boulardii, lactobacilli, passive C.
difficile antibodies, nontoxigenic C. difficile strains
for colonization, and C. difficile
vaccines, should be examined carefully in basic laboratory and clinical research studies. Until new preventive
strategies can be developed, the last sentence of a recent commentary in Lancet by Wilcox (14) seems appropriate:
"Less time spent writing antibiotic prescriptions and more time spent hand-washing would be a good starting
- Bartlett J.In: Clostridium difficile: Its role in intestinal
disease.San Diego, CA: Academic Press, Inc. 1988:1-13.
- Gerding DN, Johnson S.In: Abstracts of the First International
Conference on the Molecular Genetics and Pathogenesis of the Clostridia.Tucson, AZ, January 11-14, 1995, p. 22.
- Johnson S, Clabots CR, Linn FV, et al.Lancet 1990;336:97-100.
- Clabots CR, Johnson S, Olson MM, et al . J Infect Dis
- Gerding DN, Johnson S, Peterson LR, et al.Infect Control
Hosp Epidemiol 1995;16:459-77.
- Anand A, Bashey B, Mir T, et al.Amer J Gastroenterol
- Murphy P, LeClaire J.Abstracts of the 36th Interscience
Conference on Antimicrobial Agents and Chemotherapy.New Orleans, LA, September 15-18, 1996, p. 225.
- Hirschhorn LR, Trnka Y, Onderdonk A, et al.J Infect
- Riley TV, Cooper M, Bell B, et al.Clin Infect Dis 1995;20(Suppl
- Gerding DN.Amer J Med 1996;100:485-6.
- Katz DA, Lynch ME, Littenberg B.Amer J Med 1996;100:487-95.
- Gerding DN.Abstracts of the Anaerobe Society of the
Americas Meeting.Chicago, IL, July 19-21, 1996, Abstract 15F.
- Hospital Infection Control Practices Advisory Committee
(HICPAC).Infect Control Hosp Epidemiol 1995;16:105-13.
- Wilcox MH.Lancet 1996;348:767-8.