RESUMO
BACKGROUND: Patients with cancer are at risk of multidrug-resistant bacteria colonization, but association of colonization with in-hospital mortality and one-year survival has not been established in critically ill patients with cancer. METHODS: Using logistic and Cox-regression analyses adjusted for confounders, in adult patients admitted at intensive care unit (ICU) with active cancer, we evaluate the association of colonization by carbapenem-resistant Gram-negative bacteria or vancomycin-resistant enterococci with in-hospital mortality and one-year survival. RESULTS: We included 714 patients and among them 140 were colonized (19.6%). Colonized patients more frequently came from ward, had longer hospital length of stay before ICU admission, had unplanned ICU admission, had worse performance status, higher predicted mortality upon ICU admission, and more hematological malignancies than patients without colonization. None of the patients presented conversion of colonization to infection by the same bacteria during hospital stay, but 20.7% presented conversion to infection after hospital discharge. Colonized patients had a higher in-hospital mortality compared to patients without colonization (44.3 vs. 33.4%; p < 0.01), but adjusting for confounders, colonization was not associated with in-hospital mortality [Odds ratio = 1.03 (0.77-1.99)]. Additionally, adjusting for confounders, colonization was not associated with one-year survival [Hazard ratio = 1.10 (0.87-1.40)]. CONCLUSIONS: Adult critically ill patients with active cancer and colonized by carbapenem-resistant Gram-negative bacteria or vancomycin-resistant enterococci active cancer have a worse health status compared to patients without colonization. However, adjusting for confounders, colonization by carbapenem-resistant Gram-negative bacteria or vancomycin-resistant enterococci are not associated with in-hospital mortality and one-year survival.