The issue with blood cultures is what are you trying to diagnose.
- Bacteremia from an infection: Then you want volume. Get two sets from any place you can.Well, almost any place. Avoid the retinal vein, for example.
- Endocarditis: Then you want positive blood cultures over time. Two sets at least 4 hours apart before antibiotics.
- Catheter Infections: Then you want simultaneous cultures from the line and from the periphery. If the line cultures are positive 120 minutes before the distal, the line is the source.
Since it takes about two hours from bacteremia until fever, and because with the exception of endocarditis and line infection the bacteremia is intermittent, it should come as no surprise that it doesn't matter whether or not you get blood cultures around the time of the fever (PubMed).
We do get too many blood cultures. An isolated fever or leukocytosis has little chance of a postive culture in normal (but not immunoincompetant) patients; rigors are a much more useful as is the presense of SIRS (PubMed).
I often think the world would be a better place if more people, upon awakening in the morning, would ask themselves the simple question, "How can I make Mark Crislip's life better?" And the answer turns out to be remarkably simple. When called for positive blood cultures, before you start antibiotics, repeat the damn blood cultures. Really, how hard can that be?