Fungal peritonitis accounts for nearly 70% of invasive candidiasis in surgical resuscitation. Their mortality is high, around 38% and their morbidity also with an increase in length of stay in intensive care, invasive ventilation times and an increase in the number of surgical revisions.

Currently, diagnostic and therapeutic management is based on clinical prognostic scores, of which the "Peritonitis score" (PS) is still used as a reference. These scores are intended to best target a population eligible for probabilistic antifungal treatment pending the confirmation or not of the fungal characteristic of peritonitis by the fungal culture (Gold Standard). In case of severe peritonitis, if the PS is> or = to 3, the clinician must start a broad-spectrum antifungal treatment which will be continued if the culture of the peritoneal fluid is positive to yeast. If it is <3, it is not recommended to introduce an anti-fungal outside of very particular cases. In this situation, an anti-fungal treatment is only started if the direct examination of the peritoneal liquid made within 24 hours is positive to yeast or, failing this, to the reception of a positive culture several days later.

However, recent reviews of the "Peritonitis score" have emerged from the analysis of the "fungal peritonitis" subgroup of the AmarCand II study (underestimation in particular) , highlighting the difficulty of identifying this target population. In addition, none of these clinical scores have demonstrated superiority in guiding probabilistic antifungal therapy. Direct examination of peritoneal fluid also suffers from low sensitivity.

New diagnostic approaches using, among other things, blood markers (panfungal PCR, 1,3-Beta-D-glucan, etc.) have developed in recent years. In particular, 1,3-Beta-D-glucan has demonstrated superiority to the candida score in the early diagnosis of intra-abdominal candidiasis. [8]. His main interest lies in his strong Negative predictive value. The association of these markers is another way of working to improve their diagnostic performance. The use of these blood markers is encouraged by European and international consensus, however, their dosage in the peritoneal fluid is still under study.

One of the consequences of these diagnostic difficulties is the occurrence of inadequate prescriptions of antifungals in more than 40% of cases of fungal peritonitis in intensive care, or even more than 70% of cases in the case of any type of candidiasis. invasive.

The investigators therefore seek to study the diagnostic performance of fungal markers (BDG and PCR) increasingly realized in CHRU de Nancy in reference to the recommended tests (clinical (PS) and microbiological (ED and culture))

Inclusion Criteria:

• Any patient> 18 years old who has been treated for peritonitis in intensive care unit in CHRU de Nancy during year 2016

Exclusion Criteria:

• None

Patient> 18 years old who has been treated for peritonitis in intensive care unit in CHRU de Nancy during year 2016