Candida Endopthalmitis

We've talked about candidemia before...

One of the complications of candidemia is the development of endogenous endopthalmitis. Different rates of occular involvement are quoted from 2%-40% depending on case series.

The majority of patients will develop endopthamitis within 2 weeks of candidemia. Particularly if there is a delay in treatment of the candidemia or if the candidemia is protracted.

Early retinal exam can diagnosis, but follow up at 2 weeks is prudent as early lesions can be missed.

Treatment is covered in the IDSA guidelines (previously cited). Azole agents can be used for succeptible isolates. If there is a significant vitritis, sometimes vitrectomy with intraoccular amphotericin is required in addition to systemic therapy.

There are a number of good reviews here (treatment), here (case report and review of condition), here (BMJ review) and here (retinal lesions in sepsis, including description of occular manifestations).

Format revision and Candidemia Line Removal

Since I have left the clinical ID teaching service as a fellow -- hopefully to return as faculty in 2011, I will slightly alter the blog format.  I will highlight interesting cases that I see in my individual practice, and will highlight what I think are interesting/notable articles in the current literature.

This week's post:

Challenging dogma:  Central lines may not need to be removed early in candidemia???

Subgroup analysis of 2 RCTs in treatment of candidemia (mainly echinocandin and lipid ampho B used)
Looked at removal 24-48h vs. later removal vs. retention
842 total, 354 removed early 180 removed late 304 retained

Note that it appears in multivariate analysis that CVC removal was not associated with treatment success or mortality.  But the point estimates are in favor of removal and the CI's are wide.

In the univariate analysis CVC removal within 48h was associated with improved survival.

Why?  Were sicker patients having their lines retained leading to the perception of increased mortality (that was adjusted for in the multivariate analysis). Statistical confounding?

The associated editorial is worth reading.


I also see candidemia reported -- but not other metastatic complications such as endopthamitis, which would be clinically relevant but not noted in surveilance blood cultures.

Bottom line:  An interesting read that challenges dogma, and I think lends itself to further analysis with large enough numbers and robust enough data to exclude a meaningful clinical benefit of earlier removal.

Until then -- please remove my line promptly should I ever have a CVC related candidemia.

Candidemia

Today we discussed candidemia. The 2009 IDSA guidelines are available here.

In general:

• If the patient is critically ill, has recently been exposed to azoles, or the local prevalence of azole resistant candida is high initial therapy should be an echinocandin. Otherwise an azole like fluconazole would be appropriate.
  
• Line foci *must* be removed
• You should look for metastatic spread including the eyes or heart valves. Other investigations to look for osteomyelitis or septic thrombophlebitis should be based on history and clinical suspicion.
• Treatment duration varies depending on complication. In general, if there is no evidence of metastatic spread of infection treatment duration is 2 weeks after the last negative culture. There should be some clinical and microbiological follow-up to document relapse.
  

Candidemia redux

See previous blog. Today we saw a patient with candidemia, presumably from a urinary source. We discussed species identification, and the use of the germ tube test for rapid identification of Candida albicans.

This article discusses the clinical utility of the germ tube test, quoting a sensitivity of 87% and specificity of 100% for identifing C. albicans -- directly off of the positive blood culture! This saves ~24h in the rapid identification of C. albicans, as the germ tube is traditionally performed off a subculture on fungal medium, adding up to a 24h delay.