How would you approach treatment duration for patients with persistent Candida species fungemia with a history of a prosthetic heart valve but negative TEE/CT PET and no other identifiable source of infection?

Persistent fungemia in the presence of a prosthetic valve with no other identifiable source despite appropriate treatment is challenging. Few studies report high mortality in the nonalbicans candida group, and the risk increases if the patient is immunocompromised. A multidisciplinary management approach is critical. Is the patient a surgical candidate? Dual antifungal therapy is to be considered pending negative blood cultures. Opthalmology evaluation for endophthalmitis evaluation can be considered. If TEE/PET-CT were done early in the disease process, one can repeat the test in 1-2 weeks. Valve replacement is strongly encouraged if no source is identified. If the patient is not a surgical candidate, chronic suppressive therapy should be considered. 

Your question is all about the assessment, not the plan actually. The source is not evident, but it's present if there is persistent fungemia. Other work-ups should include ultrasounds of the limbs to look for septic thrombophlebitis (previous IV sites?) There could be a cryptic abscess that is eluding imaging. Some centers still do tagged WBC scans. Discussion with the radiologist will help determine what radionuclide they want to use. They will know the best imaging to use if they know the whole clinical picture, rather than just the few words they are typically provided. I ask them, "What imaging do you want to look at to help me figure this out?" Hardware or mesh that was not revealed in the original consult may be harboring the fungus. There may also be a contaminated batch of medicine/fluids that the patient is receiving that needs considered. Don't forget the teeth! An exercise of this type with several people - radiology, infection control, pharmacy, etc will help expand the possibilities and direct the plan accordingly. In this case, it's most likely the heart valve (common things are common) with a biofilm and so, it needs to come out. If it is the tricuspid valve, it could be left out with another placed later. If it is a left-sided valve then it would need to be replaced immediately. The heart surgeon may be reluctant unless the workup above has been carried out. A culture of the valve will verify your suspicion.