in cases of suspected SBE. If these cultures fail to retrieve the organism, then a second set of
blood cultures should be obtained between 7 and 10 days after the first. A delay of one or two
weeks in beginning treatment for subacute disease does not put the patient at risk from undue
complications. However, in patients with acute IE, antibiotic therapy must be begun within
one or two hours of the patient’s presentation How frequently antibiotic therapy suppresses
the growth of more virulent organisms such asS. aureusand gram-negatives is unknown. It is
the author’s experience that prior antibiotics have a very short-term effect, if any, on the
retrieval rate ofS. aureus. In the individual with persistently negative blood cultures but in
whom there remains a high suspicion of valvular infection, more indirect diagnostic means,
such as echocardiography, must be employed.
In the past, up to 50% of bacteria isolated in blood cultures represented contamination
(151). This figure is improving but not reaching the theoretical minimum of less than 3%. One
contaminated blood cultures may increase the total hospital bill of the patient by up to 40% by
prolonging hospitalization by four days (152–154). Obtaining only one set of blood cultures
may be worse than obtaining none at all. A single culture can neither define a contaminant or a
continuous bacteremia. Blood cultures should, at a minimum, be obtained in pairs. It is
extremely difficult to withhold treatment in an extremely ill patient with a single positive
blood culture albeit one that it is suspicious as representing contamination. Conversely, blood
cultures are often not obtained in the acutely ill individual since the patient is felt to ill to
tolerate even the slightest delay in starting therapy. In such situations it is far better to rapidly
draw at least three sets of blood cultures through separate venipunctures than not to obtain
any at all. Because the BSI of IE is continuous, there is no reason to wait to draw blood cultures
until the patient’s temperature is on the rise.
Every precaution should be taken to prevent contamination. The skin should be prepared
with 70% isopropyl alcohol followed by application of an iodophor or tincture of iodine. It
should be allowed to dry completely for maximum effect (155). Because of the risk of
contamination, cultures should never be drawn through intravascular lines except for
documenting infection of that line (156). Each set must be drawn through a different
venipuncture. Replacement of the needle before inoculating the specimen into the blood
culture bottles is unnecessary. Because of the low concentration of bacteria in most BSIs, a
10 mL aliquot should be added to each bottle to produce a 1/10 ratio of blood to broth. This
dilution may also inhibit the suppressive effect of both antibiotics and the patient’s own
antibodies (157). There is no one ideal growth medium for recovering organisms from the
blood. Trypticase soy broth is the most commonly used aerobic medium. Thioglycolate is its
anaerobic counterpart. The anticoagulant, SPS, is added to the blood culture media because
most pathogens do not thrive within blood clots, SPS also interferes with the inhibitory effects
of white cells and of several antibiotics (153). Abiotrophia spp. requires pyridoxine
supplementation for its growth. This substance is present in the broth of automated blood
culture systems. These systems make it unnecessary for cultures to be incubated for two to
three weeks for recovery of fastidious organisms (i.e., members of the HACEK group,Brucella
spp. andFrancicella tularensis). Only 50% of routine blood cultures in the setting of candidal
valvular infection are positive (47).AspergillusandHistoplasmaare rarely recovered from the
bloodstream. When specific fungal cultures are employed along with adjunctive tests
(serological), the rate of diagnosis of Candida IE may increase to 95% (158). A major
contributing factor to missing the diagnosis of fungal IE is the failure to even include it in the
differential. In one series, only 18% of the cases were suspected at the time of hospitalization
(47). This inability to recognize potential cases is increasingly more significant with the ever-
increasing numbers of immunosuppressed patients and those who are cared for in CCUs.
There are three major characteristics that the nodes each with positive culture (154):
- The type of organism recovered. CoNS that is recovered in blood cultures and
individual without intravascular catheter or other prosthetic material in place usually
represents a contaminant. - Multiple specimens that are positive for the same organism.
- The degree of severity of illness of the patient is directly proportional to the
likelihood that a blood culture result does not represent contamination.
Infective Endocarditis and Its Mimics in Critical Care 233