Les infections aiguës nécrosantes des tissus mous : importance de la localisation dans le choix de l’antibiothérapie de première intention

Abstract

Necrotizing soft tissue infections (NSTI) are infrequent but life-threatening, and require prompt empirical antibiotic therapy. Current nosologic classifications have limited value because the criteria used are imprecise and their bacteriological specificity is uncertain. The aim of this study was to describe the bacterial flora and its antibiotic sensitivity in a cohort of patients with NSTI, and to derive guidelines for the choice of antimicrobial chemotherapy. This prospective study involved 120 patients. Aerobic and anaerobic bacteriological samples were taken from infected soft tissues. The species distribution and susceptibility of the isolates to various antibiotic (ATB) combinations were analyzed. The data were analyzed according to the type (cellulitis versus myonecrosis) and anatomical location of NSTI (abdomen and perineum; uterine cervix; limbs). The chi-square test was used to analyze qualitative variables, and Student's t test was used for quantitative variables. A total of 232 samples yielded bacterial isolates (122 aerobic, 110 anaerobic). The species distribution of anaerobes did not differ according to the nature of the involved tissue or the anatomic location. Gram-negative aerobes were more frequently isolated from abdominal, perineal and limb sites than from the cervix (p<0.05), while gram-positive aerobes showed the reverse distribution (p<0.05). Metronidazole was more effective than clindamycin on cervical isolates (95% vs 88%, p=0.0093). Among the broad-spectrum antibiotics tested, imipenem/cilastatin and piperacillin/tazobactam were equally effective against the different groups of bacteria (94% vs 88%, p=0.14), and were clearly more active than the other antibiotics (p<0.05), whatever the site of isolation, the bacterial species, and the type of NSTI. The five antibiotics tested showed similar efficacy against cervical isolates. These results suggest that the choice of antibiotic therapy for NSTI should depend on the anatomical site of involvement rather than the nature of the infection. For abdominal, perineal and limb NSTI, we recommend first-line treatment with a betalactam-inhibitor combination (piperacillin/tazobactam or ticarcillin/clavulanate) plus an agent active on gram-negative species (aminoglycoside or fluoroquinolone). For cervical NSTI, we recommend penicillin G/metronidazole, or amoxicillin/clavulanic acid.