abdominal candidiasis: what positive predictive value of culture?

P. Montravers


Candida are physiological found colonizing the digestive tract in 20 to 30% of the population, even in the absence of prior antibiotic treatment. The literature reports fungal concentrations of the order of 10 2 to 10 4 cfu / ml in the stomach, intestinal or colonic. Certain circumstances, such as bariatric surgery, may increase their frequency especially in the stomach. The positive predictive value of a Yeast Culture (probability that a condition is present when a test is positive) is very difficult.

The environmental contamination and the patient's skin open drainage systems (blades, modules ...) is very common. Interpretation of the results of these crops is impossible. Experts advise against formally achieving the drainage samples. The recommendations are not to be considered as deep samples, intraoperative surgical or percutaneous drainage during interventional radiology procedures.

Very little is available to enhance the reflection on the value of a positive sample for culture. Biomarkers of inflammation such as procalcitonin and C-reactive protein are of little help. No studies have focused on the mannan antigen / anti-mannan antibodies in peritonitis. ß-D-glucan has been little studied, but may help in interpreting the results. Studies of PCR are also very limited.

The clinic needed to analyze two different situations: the Community infections and healthcare-associated infections.

During community infections, candida positive samples are reported in 10-20% of cases, mostly in gastroduodenal or colonic perforation. These samples are mostly multi-microbial, combining the Candida aerobic and anaerobic bacteria in the digestive flora Gram-positive and Gram-negative. The frequency of candidemia is very low.

Data from the literature suggest that it is not necessary to establish a probabilistic antifungal treatment for Community peritonitis in the absence of signs of seriousness, except in immunocompromised patients, transplant or with an inflammatory disease of the tube digestive. In case of septic shock, clinicians frequently introduce an antifungal empirical treatment. At the stage of final results, an antifungal treatment is maintained in a limited number of cases.

We can conclude that in Community peritonitis, except in patients at risk or septic shock, a positive culture for Candida is often regarded by clinicians as a reflection of colonization.

In peritonitis associated with care, post-operative infections are the most common and most studied. No study has ever specifically assessed value of microbiological tests in these conditions. Compared with community infections, the frequency of positive samples candida is higher in the order of 20 to 30% of cases. These levies are frequently multi-microbial, but pure fungal samples are reported in 10-20% of cases. In patients reoperated after a complication of bariatric surgery, yeast isolation frequency is high. Finally, candidemia were observed in 5 to 10% of cases.

During the postoperative infections, isolation of yeasts in peritoneal fluid culture is associated with excess mortality. This excess mortality was also found in a case-control study in patients with postoperative peritonitis.Based on these studies, experts suggest starting empiric antifungal therapy in patients with severe sepsis or septic shock in care associated infection box. The therapeutic adaptation after culture result is also recommended.

Overall, among patients with health care-associated infection, a positive culture for Candida is considered a reflection of a genuine infection. This attitude may well be excessive. This therapeutic choice based on an extrapolation of work performed in candidemia that show an increase in mortality in parallel within the establishment of effective antifungal therapy. No such study has been conducted in intra-abdominal infections and evidence of this argument lacking.

Ultimately, only the therapeutic test can remember a pathogen or no value to a candida isolated from a surgical specimen. A study on a series of peritonitis Candida chose this reasoning dealing initially bacteria in surgical specimens and then adding an antifungal if patients continued to show clinical signs of infection or failure to improve.


Full conference title: 

Réunion Interdisciplinaire de Chimiothérapie Anti Infectieuse
    • RICAI 34th (2014)