RESUMO
Background: Empirical antibiotherapy (EA) should target all bacteria in post-operative peritonitis (PP). Nevertheless, recent studies failed to prove a link between adequacy of EA and prognosis of PP. We sought to confirm this loss of association between adequate EA and prognosis and to analyze the evolution of patients' characteristics and antimicrobial strategies. Methods: This is was retrospective study. Patients with a positive fungal culture were excluded. The cohort was divided into two time periods. Data of survivors and non-survivors were compared within each time period. Differences between the two periods were assessed. A multivariable analysis searched for parameters associated with a higher hospital mortality rate. Results: Two hundred fifty-one patients were included, with 92 patients in the first period (P1) and 152 patients in the second period (P2). Inadequate EA was associated with a worse outcome only in P1. The multivariable analysis in the whole cohort showed that inadequate EA was associated with a higher mortality rate. When the differences noticed between the two periods were entered in the model (presence of resistant gram-positive cocci and EA comprising glycopeptides), inadequate EA was no longer associated with worse outcome. In P1, the most severe patients had more resistant bacteria, hence, had a higher rate of inadequate EA. This artifact disappeared in P2, during which broader antibiotherapies with triple EA were more often prescribed for the most severe patients. Conclusion: This study showed that the link between inadequate EA and outcome of patients with PP was at least partly artifactual in older studies.
