Introduction
The Clostridium difficile
bacterium is a gram positive rod. It exists in a vegetative and spore
state and some strains produce toxins. The first reports of
Clostridium difficile being isolated are from the 1930’s, however,
it was not linked to significant morbidity until the late 1970s. At
that point, it was linked to pseudomembranous colitis and diarrheal
illness in patients on antibiotics. From the 1970s until early in the
2000s, the organism was seen as a problem in hospitals and skilled
nursing facilities but not a major cause of mortality in large
numbers. Outbreaks were reported demonstrating that the organism could
be transmitted from person to person and various antibiotics were
identified as risk factors for developing illness. The epidemiology
and profile of Clostridium difficile has dramatically changed in the
last two years.
Clostridium difficile may or
may not cause disease in specific individuals. It has been found as
intestinal flora in up to 70% of newborns while causing no acute
illness. C. difficile has also been identified in a much smaller
percent of healthy adults while also not causing any clinical illness.
This in part has to do with whether the Clostridium difficile
identified is a toxin producing strain. Only toxin producing strains
are considered pathogenic. It should also be noted that the presence
of a toxin producing strain in the intestines does not assure that a
patient will develop symptoms. In fact, up to 30% of hospitalized
patients who receive antibiotics may remain asymptomatic carriers.
Clostridium difficile may produce three toxins, TcdA, TcdB and CDT.
Both TcdA (or Toxin A) and TcdB (or Toxin B) are capable of causing
cell death and colonic dysfunction. Toxin A is the most common toxin
produced by C. difficile. CDT is unique in that it is a binary toxin.
In cell cultures CDT has been found to cause cell death. Its role in
causing disease in humans is unclear as most cases of binary toxin
producing Clostridium difficile also produce Toxin A and/or Toxin B.
However, there has been at least one report of a strain of Clostridium
difficile not producing Toxin A or B and producing the binary toxin
being isolated from a patient with diarrhea. Binary toxin producing
strains can be found in 1.6% to 20.8% of C. difficile clinical
isolates. In some outbreak settings, binary toxin producing strains
have become the predominate strain.
The spectrum of disease caused
by the toxigenic strains of Clostridium difficile can range from a
mild diarrhea to more severe pseudomembranous colitis and toxic
megacolon, both of which can result in death.
Diarrhea is the most common
manifestation of Clostridium difficile associated disease (CDAD).
However, it is incorrect to think of CDAD as only a diarrheal disease
since solid stool with systemic symptoms can be a less common
manifestation of CDAD. Patients with non-bloody diarrhea sometimes
accompanied with mild cramping and mild abdominal tenderness are
epidemiologically described as having mild disease.
Moderate disease involves
profuse diarrhea, fever, abdominal pain and leukocytosis.
Pseudomembranous colitis is a
much more severe manifestation of Clostridium difficile. The diarrhea
is profuse and watery and the patient usually presents with abdominal
distension and pain. In stage 4 pseudomembranous colitis, the patient
will have bacterial overgrowth of the walls of the intestine and the
typical light yellow plaques found in Stages I-III of pseudomembranous
colitis will not be able to be visualized on endoscope examination.
While the patient may be positive for occult blood in the stool, frank
blood is rare.
Patients with severe colitis
can progress onto paralytic ileus and toxic megacolon. Development of
megacolon will stop the patient’s diarrhea as the peristaltic action
in the part of the intestine affected with megacolon will stop and the
megacolon will fill with waste material. The patient in these cases
will present with an acute abdomen, fever and tachycardia.
Other manifestations of severe
CDAD include sepsis, peritonitis, hypotension, volume depletion,
electrolyte imbalance, elevated creatinine and a white blood cell
count greater than 20x109/L.
Recent data suggest that the
patient presentation may depend on whether they have a strain that
produces the binary toxin. Patients who have binary toxin producing
Clostridium difficile were significantly more likely to have shorter
duration of diarrhea. In that study the small sample size limited
findings, however two associations, while not significant, were noted.
Abdominal pain occurred in 63.6% of the cases with binary toxin
strains compared to only 39.4% among strains without binary toxin
production and 23.1% of the patients with a binary toxin producing
strain had normal stools while 40.5% of the patients without the
binary toxin strains had formed stools. The fact that around 25% or
more of individuals who have Clostridium difficile associated illness
did not have diarrhea underlines the problem of diagnosis and of the
spectrum of disease being commonly called Clostridium difficile
associated diarrhea.
The above description is how
the disease was known to clinicians through the late 1970’s until
the early 2000’s. But in the early 21st century two outbreaks, one
in Pittsburgh, Pennsylvania and the other in Quebec, Canada noted
increased morbidity and mortality in the patient population suffering
from C. difficile associated disease. The strain responsible for these
outbreaks (called BI/NAP1 strain) was noted to have a specific
mutation (an 18-bp tcdC deletion). In layman’s terms, this mutation
eliminated the down regulation for Toxins A and B. Instead of the
bacteria limiting production of Toxin A and B as it normally does,
this mutated bacteria was constantly producing Toxins A and B. The
increase in mortality and morbidity was dramatic. In the Quebec
outbreak, the number of deaths were reported to be as high as 2000 at
the peak of the epidemic. This number is in dispute by some officials.
Although a 30-40% drop in cases has been noted in Quebec, the number
of deaths among 5113 patients with Clostridium difficile associated
disease from August 2004 to August 2005 is acknowledged to be 409.
Deaths were reported to occur more frequently among the elderly.
Diagnosis
There are a number of
different mechanisms to test for Clostridium difficile. , , , In the
past, the most common test was the anaerobic stool cultures. In recent
years it has fallen into disfavor. This test sensitivity of the test
is good (89-100%), but anaerobic stool cultures can take approximately
72 hours to provide results. This test also has a very high false
positive rate because the test cannot differentiate between toxigenic
and non-toxigenic strains.
For all stool-based testing,
laboratories should only accept watery or loose stools when evaluating
patients for clinical illness. Given the rates of colonization in
hospitalized patients, the presence of toxin in formed stool in
asymptomatic patients proves only colonization and not disease. No
laboratory test currently exists to evaluate whether treatment of CDAD
was efficacious. Despite treatment success as measured by clinical
improvement, patients can continue to test positive.
.Endoscopic evaluation is also
used for evaluation. Its sensitivity is far lower (~51%) but in some
cases it may be the best tool. Pseudomembranous colitis is
pathognomonic for C. difficile and is best identified via endoscopic
examination or examination of pathology samples.
Tissue cytoxic assay is
probably the gold standard as it exists today for accurate diagnosis
of C. difficile related illness. Results are dependant on the skill of
the technologist running the test. It can detect most pathogenic
strains but can have false positive results and results are not
available for about 2 days.
Antigen testing is also
available for C. difficile. This test has the shortest turnaround
time, between 15-45 minutes, but its sensitivity is as low as 58% and
cannot differentiate between toxigenic and nontoxigenic species of
Clostridium difficile. There have also been reports of cross
reactivity of the antigens to other anaerobes thus making the test
difficult to interpret.
Immunochromatographic toxin A
testing traditionally has results in less than 1 hour but the
sensitivity ranges between 60%-85% and only detects toxin A, not
strains that may produce only toxin B or only binary toxin.
Two enzyme linked
immunosorbent assays (ELISA or EIA) for Clostridium difficile toxin
are available. They are not equal. One EIA test only detects toxin A
and is easy to use with about a 2 hour turnaround from sample to
result. Sensitivities reported vary but are between 70%-95%. The
second EIA test detects both A and/or B toxin. The sensitivities are
the same as are the turnaround of the test. This is the most widely
used test for Clostridium difficile toxin detection in the U.S. Feces
should be submitted within 2 hours of the specimen collection or
should be refrigerated.
The low sensitivities of the
ELISA tests have resulted in some practitioners repeating the test. A
recent study has shown that the yield on repeat testing is extremely
low and probably not worth the additional cost.
As discussed above, all the
commercially available tests currently in use have limitations. None
have the ability to detect binary toxin although research laboratories
have developed methods not commercially available. They either have
false positives or take long periods of time to produce results for a
disease that can be rapidly progressing. Currently, commercial PCR
tests are under development. Whether these tests will be an
improvement or just present a new set of limitations has yet to be
seen.
Risk factors
The most widely identified
risk factor for CDAD is antibiotic use. Antibiotics disrupt the normal
intestinal flora removing other competing bacteria. This allows
Clostridium difficile to proliferate. Clindamycin was among the early
antibiotics associated with Clostridium difficile related disease.
However, almost all antibiotics except for aminoglycosides (i.e.
amikacin, tobramycin, gentamicin) have been linked in one study or
another with CDAD. This includes, ironically in some univariate
analysis, vancomycin, an antibiotic used to treat CDAD. Most experts
feel that the broader spectrum the antibiotic the greater the
likelihood for disruption of the intestinal flora and the greater the
likelihood of development of disease if the patient has the bacteria.
Recent reports about the new strain of C. difficile responsible for
the Quebec and Pittsburgh outbreaks have noted fluoroquinolones,
gatifloxacin and moxifloxacin in particular, as a risk factor for
developing illness caused by the BI/NAP1 strain. This strain is in
fact resistant to both of those fluoroquinolones, a characteristic
unique to it among Clostridium difficile strains.
Other risk factors identified
in the literature are age greater than 65 years, severity of
underlying illness, nasogastric intubation, being in a healthcare
facility (acute or long-term) , and being in a healthcare facility
with an outbreak of Clostridium difficile associated disease. The last
risk factor is important in association to what antibiotics are risk
factors for CDAD. The outbreak literature shows that the antibiotic
putting the patient at increased risk is specific to that event. This
may be because certain strains of Clostridium difficile are able to
succeed in unique niches caused by the antibiotic milieu specific to
that facility during a limited time period.
A major controversy has
developed in the field over whether proton pump inhibitors used to
suppress gastric acid production are a risk factor for development of
CDAD. The alteration of gastrointestinal flora by these antacids could
be the causative factor if indeed these drugs are a risk factor.
Coincidently, it has been noted by some that gastric bypass surgical
patients during their recovery phase also appear to be at increased
risk for CDAD although this has not been reported in the literature.
Whether having surgery or being a medical patient appears to place one
at greater risk for developing CDAD has been studied and contradictory
findings have been reported. , The answer to this question may depend
which population at a specific institution receives more antibiotics
and what other risk factors each population has.
Two major protective factors
have been identified in the literature against developing CDAD. These
are youth and duration of colonization with Clostridium difficile. As
previously mentioned, up to 70% of neonates are noted to be colonized
with strains of C. difficile. While not all of these are toxigenic,
some are. Animal models suggest that the lack of illness caused by
these toxigenic strains may be due to the fact that neonates lack the
receptor site to uptake toxin A in their enterocytes. However, if this
is the sole reason C. difficile is not reported as causing illness in
neonates, then strains of Clostridium difficile that do not produce
toxin A but produce either toxin B and/or binary toxin could be
pathogenic to neonates.
Lastly, prolonged colonization
with Clostridium difficile has been identified as a protective factor
against developing CDAD. This protective factor of prolonged
colonization has been linked to the development of antibodies to toxin
A. As discussed above, colonization rates in healthy adults are about
3%, but in patients in long-term care facilities rates have been
reported to be anywhere from a comparable 4% to as high as 20%.
Preventing the occurrence of additional cases
of CDAD
As previously discussed, the
impact of CDAD on the health of a community has ranged from
spontaneous isolated cases to impacting single facilities all the way
up to whole cities and regions. Although the literature does contain
interruption of outbreaks based on focusing on one approach, ideally,
control efforts need to be three pronged. The first prong is
elimination of transmission through improved healthcare worker hand
hygiene, the second is elimination of risk factors for development of
CDAD, specifically restriction of the antibiotic linked to the
outbreak, and the third prong is elimination of reservoirs for
organism, more specifically, removing the organism from the
environment.
Most of the transmission of C.
difficile probably occurs on the hands of healthcare workers carrying
the bacteria contained in one patient’s feces to the second patient’s
environment where it is ultimately ingested. Therefore, the CDC
recommends for healthcare workers practice strict hand hygiene and
place patients in contact precautions if they are identified as having
CDAD. However, recent CDC guidelines may not have been so helpful at
containing Clostridium difficile. The CDC guideline on hand hygiene
launched a major shift in hospitals away from hand washing which is
difficult because of inadequate numbers and poor locations of sinks in
facilities to the wide spread use of alcohol based hand rubs. While
this resulted in a marked improvement in the number of times hand
hygiene was practiced, its implications for the control of CDAD were
less clear. C. difficile is a spore forming organism. In its
vegetative state, C. difficile can be killed by alcohol but in its
spore state it remains viable. Some institutions have switched away
from alcohol based hand rub use in cases of patients with known CDAD
and are using soap and water exclusively. However, in such
institutions, given the colonization rates reported of asymptomatic
patients with C. difficile who could be potential reservoirs, it is
unclear whether such an approach would be entirely effective at
eliminating transmission. Fortunately though, the bioburden of
Clostridium difficile in the rooms of colonized patients has been
found to be far lower than that found in patients with CDAD. While
aggressive, active surveillance for those who are colonized with MRSA
and VRE has been suggested as a way to control transmission of these
pathogens, to date no professional organizations have advocated active
surveillance and placing patients in contact precautions for
colonization with Clostridium difficile. Nor is it recommended to
treat colonization with Clostridium difficile as an infection control
measure.
There are reports of outbreaks
being interrupted by severe restrictions on the antibiotic implicated
as the major risk factor for development of CDAD. However, many other
outbreaks have continued after draconian restrictions of the
implicated antibiotic suggesting that such restrictions alone are
frequently not enough to stop an outbreak. Nevertheless, once an
outbreak is identified among the first tasks that an institution
should attempt is identification of the antibiotic that is increasing
the risk for CDAD and restricting its use.
The environment has been shown
to be a reservoir for Clostridium difficile. Traditional cleaning of
the patient’s room is ineffective since most organizations do not
use dilute bleach solutions for such tasks. Bleach has been
recommended for cleaning the rooms of patients with CDAD in outbreak
settings and has been shown to reduce CDAD rates in some settings. ,
But given the limitations of bleach (its ability to pit the
environment, and respiratory issues when used over large surfaces), it
can not be recommended for cleaning of the entire healthcare
environment. Ironically, since bleach is a respiratory irritant,
patients at greatest risk for acquisition of C. difficile probably do
not have their environment cleaned with bleach. A case in point is a
patient who shares a room with another patient who is incubating CDAD.
Traditionally in hospitals with semi-private rooms, when one of the
two patients develops CDAD, the patient with CDAD is moved to a
private room. Due of the scarcity of rooms, a new patient is moved
into the CDAD patient’s old room before any terminal cleaning with
bleach occurs. Given its ability to irritate the respiratory system,
use of bleach is rarely conducted in occupied rooms. In the case
described above, if the environment plays a critical role in
acquisition of Clostridium difficile, two patients (the original
roommate and the person subsequently occupying the CDAD patient’s
place in the room) have been placed at elevated risk of acquiring
CDAD.
Patient care equipment has
been implicated in the transmission of C. difficile. In the case of
contact precautions, equipment must be dedicated solely to one patient
or if not possible the equipment must be cleaned between patients. One
interesting fact is the role, or more correctly, lack of role of
endoscopes and their disinfection. For most applications, endoscopes
are usually high level disinfected rather than sterilized. High level
disinfection kills Clostridium difficile in its vegetative form but
not in its spore form. A review of the literature and press reports
show scopes have been found to be inadequately cleaned or disinfected.
Despite this, endoscopes have not been reported to have played a
prominent role in CDAD outbreaks.
The most promising strategy to
combat the spread of CDAD is in the research and development of an
effective vaccine. Phase I clinical trials are currently underway to
test the efficacy of a vaccine in elderly patients as well as those
with recurrent or relapsing CDAD. The vaccine has already demonstrated
to be highly immunogenic in healthy human volunteers. However it is
unclear who will be the target population for this vaccine if it is
released, given CDADs’ spread into the community.
One method of controlling the
spread of CDAD that has been shown to be ineffective is the use of
probiotics. Probiotics in this case are benign bacteria that compete
for the same niche as Clostridium difficile in the patient’s enteric
flora. The approach had been to introduce these to patients already
exposed to C. difficile or known to be at risk but without current
disease. Unfortunately, no significant difference was found in CDAD
attack rates between those who received probiotics and those who did
not.
Treatment
Currently there is no
consensus on the treatment algorithm for CDAD. Some basic rules for
treatment exist for which there is universal consensus.
The first rule of successful
treatment of CDAD is that the antibiotic that was associated with the
CDAD should be stopped. In many cases this is simple as some cases of
CDAD can occur even after the initiating antibiotic has been
discontinued for over a month. Nevertheless, continuation of the
antibiotic has been linked to treatment failure and inferior patient
outcomes. This can be challenging in cases where the antibiotics were
initiated for conditions associated with high mortality and due to
drug resistance, where antibiotic alternatives are extremely limited
or non-existent. In these cases, because of BI/NAP1 is associated with
rapid progression and high mortality, the decision to discontinue the
inciting antibiotic can be very difficult. Before the emergence of
BI/NAP1, stopping the inciting antibiotics was shown to have moderate
success on its own in treating CDAD.
While traditional treatment of
diarrheal illness includes antiperistaltics, in cases of CDAD,
antiperistaltics can be deadly. This increased risk of mortality
associated with antiperistaltic agents in patient’s with CDAD is
easy to understand. Peristaltic action reduces the amount of time the
toxin is in contact with the intestinal lining, thereby reducing the
damage it does to those cells. Slowing peristaltic action extends
contact time with the toxin which worsens disease. In patients with
CDAD, all drugs that the patient is on should be reviewed for their
impact on peristaltic action as other drugs such as opiates have
antiperistaltic properties and may place the patient at risk for a
poorer outcome.
Given the limitations of the
current methods of testing for CDAD, there could be the understandable
belief that empiric treatment for CDAD would be necessary. Currently,
recommendations for mild diarrheal illness not known to be CDAD do not
include empiric treatment. This is because even in outbreak situations
only 30% of cases of antibiotic associated diarrhea are thought to be
due to Clostridium difficile.
Up until 1995, treatment of
CDAD was driven by the physician’s personal preference. Studies had
shown that metronidazole was equivalent to oral vancomycin (IV
vancomycin does not reach the intestinal system in therapeutic levels)
in treating CDAD. But in 1995, in response to growing concerns over
the rapid increase in vancomycin resistant enterococcus, the CDC
released recommendations to treat CDAD with metronidazole. Ten days of
metronidazole therapy became the standard and if the patient failed
metronidazole therapy, the patient was then given oral vancomycin
therapy of an equal duration.
The issue of front line
therapy for CDAD has been revisited because of the emergence of the
BI/NAP1 strain, with its reports of increased of treatment failure
with metronidazole and also increased rates of relapse. . Since the
data showed lower success rates treating CDAD with metronidazole than
had been previously published there were legitimate concerns that
Clostridium difficile was developing resistance to metronidazole.
However, studies have not found high rates or increasing rates of
resistance to metronidazole. Most likely the initial treatment
failures reported are due to poor fecal concentration of metronidazole
in patients with no diarrhea and suboptimal host immune response.
The issue of recurrence of
CDAD is more complex. Between 12-24% of patients have a second episode
of CDAD within 2 months of their first episode. The risk of recurrence
of another episode goes up with each event. Recurrence of CDAD has
three plausible explanations: treatment failure to eliminate
vegetative Clostridium difficile in the patient’s intestines, or
re-infection of the patient with Clostridium difficile, or spore
recrudescence. It has been demonstrated that approximately half of all
of patients thought to have had a relapse of CDAD within 2 months were
actually infected with a new strain. Reinfection may occur more
frequently than 50% because the patient’s own environment is most
likely contaminated with their own strain and there is no test for
re-infection with the patient’s own strain from their external
environment. Because of fear of recrudescence some physicians have
extended therapy. This is generally not recommended. Altogether
though, the relapse of CDAD is too complex and poorly understood to
solely base any wholesale change on what the first line therapeutic
agent of choice is.
Out of these events have
emerged divergent practice choices for treating mild CDAD. The first
and least supported by the literature is the complete switch from oral
metronidazole to oral vancomycin. Adherents to this school support
their position by citing the reports of treatment failures with
metronidazole and that rates of treatment success are lower than was
originally reported when the CDC advocated abandoning oral vancomycin
as a first line treatment for CDAD.
Others argue that the
treatment failures with metronidazole were due more to host factors,
that metronidazole resistance is still very rare and that oral
vancomycin had it been the first line drug for treatment of CDAD would
be suffering the same treatment failure record.
A more nuanced approach
currently has been developed for treatment of CDAD. For cases of mild
CDAD, metronidazole is still recommended although vigilant monitoring
for response to therapy is recommended as CDAD can progress rapidly.
For cases of moderate or
severe disease some have suggested using oral vancomycin instead of
metronidazole. It should be noted that there are no blinded randomized
prospective studies that have shown that this approach is superior.
However, if the disease has progressed from mild disease to moderate
disease while receiving metronidazole, switching to oral vancomycin
would appear to be prudent.
Clinical response to therapy
is generally seen as a discontinuation of diarrhea in 2-5 days. In
febrile patients, the fever responds more rapidly, usually in 1-2
days. If the patient deteriorates or symptoms do not lessen after 6 or
7 days of therapy, a surgical consult should be obtained for
consideration of possible colectomy. For these patients, intracolonic
vancomycin may also be considered.
In cases where patients are
unable to take oral medications, intracolonic vancomycin has been
shown to be effective and is the first choice. Metronidazole in the IV
form has not been tested with measures for patient outcomes but it has
been found to produce therapeutic concentrations in the feces.
Besides the above therapies,
more novel approaches have been suggested. A review of alternative
treatment modalities for CDAD found that in preventing recurrences
three options showed some promise: tapered dose vancomycin,
lactobacillus, and fecal enemas all demonstrated a reduction of in
recurrence rates of CDAD.
Myers is manager of clinical
epidemiology and safety systems at Scripps Mercy Hospital in San
Diego, Calif.
