» How do you tell when
a positive blood culture is significant?
One can never assume that a blood culture is a contaminant
without first carefully considering the patient and the organism.
Common isolates that always or near always (>90%) represent true
infection include S.aureus, E.coli and other Enterobacteriaceae,
P.aeruginosa, S.pneumoniae, and Candida albicans. Other organisms
such as Corynebacterium spp., Bacillus spp., and Propionibacterium
acnes rarely (<5%) represent true bacteremia (1). Some organisms,
particularly S.epidermidis, are more problematic in that they
often represent contamination although in the appropriate patient
and circumstance, they can be pathogenic.
S.epidermidis is a coagulase-negative staphylococcus
(CNS) that is part of the normal skin flora and of lower virulence
than S.aureus. The proportion of all bloodstream infections caused
by CNS is increasing (2). It is the most common organism isolated
accounting for approximately 40% of all positive blood cultures
(1). Forty-one percent of the positive blood cultures from our
institution to date this year have been secondary to CNS. It is
also the most common contaminant, accounting for 75% of contaminated
blood cultures (2). While most of these blood cultures represent
contamination approximately 12-15% are clinically significant
(1,3). Fortunately, nosocomial bacteremia secondary to CNS is
associated with a significantly lower risk of death than other
nosocomial bloodstream pathogens after adjusting for confounders
(2). The attributable mortality is approximately 5% (1,2).
There are certain factors that help to determine whether
CNS in a blood culture represents true bacteremia or a contaminant.
CNS has certain adherence factors as well as the ability to produce
an extracellular glycocalyx, both of which make it a common cause
of true bacteremia in those patients with indwelling prosthetic
material. It is therefore the most common causative organism in
intravascular-catheter-related infections, early prosthetic valve
endocarditis, and early prosthetic joint infection. Weinstein
and colleagues (1) have found that intravascular catheters as
the source of bacteremia has increased from 3-19% of cases in
the last 20 years. This in large part explains the increase in
CNS bacteremia. There is no difference in the positive predictive
value of one versus two blood cultures positive for CNS in predicting
true bacteremia (4). However, the number of positive blood cultures
out of the total cultures obtained may be helpful.
There are certain issues regarding the acquisition
of blood cultures that can decrease the percent of blood cultures
that are contaminants. Skin antisepsis is paramount and should
include 70% isopropyl alcohol, followed by an iodophor or iodine
tincture. Blood culture teams consistently do better than house
staff. Although somewhat controversial venipuncture results in
less contamination compared with cultures obtained from intravenous
catheters (3).
References:
- Clin Infect Dis 1997;24:584-602.
- Infect Cont and Hosp Epi 1998;19(8):581-9.
- Clin Inf Dis 1996;23:40-6.
- Clin Inf Dis 1996;22:14-20.
Response by: David Richardson, M.D.
Infectious Diseases Fellow
University of Toronto
» What is the significance
of Candida in a urine specimen?
Candida in the urine frequently represents colonization particularly
in association with indwelling urinary catheters. True infection
is uncommon compared with colonization. The prevalence of candiduria
has been estimated at 0.2%-6% among asymptomatic volunteers and
at 6.5%-20% among hospitalized individuals (1). The presence of
risk factors such as urinary catheters, anatomic or functional
abnormalities of the genitourinary (GU) tract, exposure to antibacterial
agents, immunosuppression and/or leukopenia, diabetes mellitus,
and candida cultures positive from other sites all lend support
to the clinical diagnosis of candidal urinary tract infection
(2).
Fisher and colleagues (3) have suggested an algorithm
for the evaluation of candiduria based on the sparse literature
in this area. They recommend first verifying the presence of the
organism by repeating a urine microscopy and culture. Of course,
the urine should be carefully collected with a clean-voided midstream
urine sample. If candida is absent with the second urine collection
it means that the first positive sample represented either contamination
or self-limited candiduria. The physician should then consider
the patient and risk factors as outlined above and stratify them
into one of three categories. The first category includes previously
healthy individuals with asymptomatic candiduria. It is well established
that these patients have favorable outcomes without treatment
(3). These patients however should be evaluated at least for diabetes,
renal insufficiency, and anatomic GU abnormalities. The second
category includes those predisposed to candiduria because of their
risk factors but in whom disseminated candidiasis is unlikely.
The third category includes those predisposed to candiduria but
in whom disseminated candidiasis is likely. The presence of disseminated
candidiasis may be determined by examining the optic fundi and
skin for candidal lesions, inspection of vascular-access sites,
and obtaining blood for fungal cultures. Patients who are unstable
or deteriorating and in whom disseminated candidiasis is likely
should be treated empirically even if no direct evidence of dissemination
is available. For all patients, predisposing conditions should
be eliminated and/or treated. The authors suggest that all patients
with candiduria who have not had their GU tract instrumented (ie.
bladder catheterization) should have their GU tract investigated.
Whether or not to treat with antifungal agents is fairly clear
for patients who are asymptomatic and those who are critically
ill, but for those inbetween it is not clear-cut. In general,
all symptomatic patients should be treated, as should those with
urinary tract obstruction. The type and duration of therapy should
take into account the presence or absence of obstructing fungus
balls, the anatomy of the collecting system, and the presumed
site and extent of infection.
This algorithm has overall been supported by the literature.
In 26 patients with confirmed candidemia and a urinary tract source,
23 (88%) had a urinary tract abnormality and 19 (73%) had urinary
tract obstruction (1). In 1997, the results of an international
conference for the development of a consensus on the management
and prevention of severe candidal infections were published (4).
They discussed several issues in relation to candiduria. There
was consensus that patients without diabetes mellitus, GU abnormalities,
or a renal transplant with asymptomatic candiduria should not
receive treatment. Their rational was that candiduria is often
caused by bladder catheterization or antibiotic use and that removing
these factors usually leads to clearing of the candiduria. They
recommended that all patients with candiduria and prosthetic materials
in the urinary tract should have these removed if possible. Consensus
was also reached that patients with candiduria who were to undergo
a GU tract procedure be treated.
References:
- Clin Infect Dis 1993;17:662-6.
- Medicine North America 1995; June:519-29.
- Clin Infect Dis 1995;20:183-9.
- Clin Infect Dis 1997;25:43-59.
Response by: David Richardson, M.D.
Infectious Diseases Fellow
University of Toronto