Falsely negative blood cultures currently occur in 5% cases of IE. These are most
frequently due to the prior administration of antibiotics (159), ranging from 35% to 79% of false
negative cultures. The false negative rate is directly related to the frequency of fastidious
organisms of (i.e.,Bartonellaspp) in the environment. This figure is most likely higher for
patients in CCU because of the multiple courses of antibiotics that are empirically given to treat
fevers that are in reality a result of undiagnosed valvular infection. This produces the state of
“muted” IE in which the valvular infection goes on while the blood cultures remain negative.
Paizin provides a specific example of this phenomenon (160). He demonstrated that the
recovery rate of streptococci from blood cultures in patients who had received any antibiotic in
the previous two weeks was reduced to 64% is compared with 100% of those patients who had
not been given antibiotics. The shorter the course of the antibiotic, the shorter the time it takes
the blood cultures to become positive. If the prior course of antibiotics has been prolonged,
then it may take up to two weeks of being off of them to be able to detect the pathogen. In the
author’s experience, antibiotics to be at the suppressive, if at all, the retrieval ofS. aureusfor a
few days only (161). Broth may be supplemented with not only sulfopolyanetholsulfonate
(SPS) but also resins (BACTEC resin) (162) that theoretically will inactivate whatever
antibiotics may be present. This approach has had a moderate amount of success in cases ofS.
aureusBSI and fungemia (163).
In the author’s experience, the second most common cause of false negative blood
cultures, especially in CCU IE, is produced by a surface sterilization phenomenon. For
unknown reasons, the infecting organisms, especiallyS. aureus, leave the surface of the
vegetation and penetrate deep within. The BSI stops but the bacteria continue to replicate and
to burrow the base of the valve. Paravalvular and/or septal abscesses and ruptured chordae
tendinae may be the final result of this process (164). Surface sterilization is most likely
becoming more frequent because of the rise inS. aureusIE.
Because of the risk of contamination, blood cultures should never be drawn through
intravascular lines except for the purpose of documenting line infection. The traditional
approach has been the role plate method. This necessitates that the catheter be removed. Only
its external surface is cultured. Approximately 80% of intravascular catheters that have been
removed because of clinical suspicion of infection have been found to be not infected. Clearly,
methods that can diagnose a CRBSI while the catheter is in place are more desirable (165).
Paired quantitative blood cultures, drawn through the catheter and peripherally, appear to be
the most accurate way to diagnose CRBSI (166). However this technique is expensive and
labor-intensive with opportunities for contamination. The differential in time of growth
between blood cultures drawn through the intravascular lines and those drawn peripherally is
much more practical way to assess the CRBSI. It makes use of the fact that automatic blood
cultures systems continuously monitor for and record the time of initial growth. The blood
culture, obtained from the intravascular device, becoming positive more than two hours
before, which obtained peripherally, reflects a heavier bacterial growth in the catheter. This
would indicate that the intravascular catheter is the source of the BSI. Semiquantitative
cultures from the hub and skin (superficial cultures) that grew out the same organism is
isolated in a venous blood culture provided approximately the same sensitivity and specificity
of diagnosing CRBSI as the preceding two methods (167).
The question of how many blood cultures are necessary to diagnose a BSI in the era of
automated blood culture systems. In a recent study, one to four sets blood cultures detected
cumulatively 73.1%, 87.7%, 96.9%, and 99.7%, respectively. Three sets are the probable optimum
number since the difference in yield is essentially insignificant between three and four blood
cultures with the possibility of increased contamination as more cultures are drawn (168).
Diagnosis of IE that is caused by pathogens that are challenging to culture in the clinical
microbiology laboratory (e.g.,C. burnetii,Legionella) is dependent on the use of serologic
studies and various types of DNA amplification techniques (169–171). PCR techniques have
been applied directly to explanted valvular tissue obtained at surgery. Limited experience
indicates that they are more sensitive and from more specific than standard cultures that have
a high rate of contamination (172).
Abnormalities of cardiac conduction are seen in 9% of patients with valvular infection.
These are due to septal abscesses or myocarditis (173). During the first two weeks of treatment
234 Brusch