The 30-day mortality was higher in the Gram-negative group
(42.1% vs. 30.8%) and amputations were more frequently
performed in Gram-positive patients (11.5% vs. 5.3%). Very few
amputations, however, were performed. Mortality rates were
higher among patients given appropriate empirical treatment
compared with those treated with inappropriate therapy. It is
possible that the more severely ill patients were those receiving
broad-spectrum antibiotics, and thus treated appropriately but
nevertheless died