Monotherapy versus beta-lactam-aminoglycoside combination treatment for gram-negative bacteremia: a prospective,...

The aim of the present study was to test whether the combination of a beta-lactam drug plus an aminoglycoside has advantage over monotherapy for severe gram-negative infections. Of 2,124 patients with gram-negative bacteremia surveyed prospectively, 670 were given inappropriate empirical antibiotic treatment and the mortality rate in this group was 34%, whereas the mortality rate was 18% for 1,454 patients given appropriate empirical antibiotic treatment (P = 0.0001). The mortality rates for patients given appropriate empirical antibiotic treatment were 17% for 789 patients given a single beta-lactam drug, 19% for 327 patients given combination treatment, 24% for 249 patients given a single aminoglycoside, and 29% for 89 patients given other antibiotics (P = 0.0001). When patients were stratified according to risk factors for mortality other than antibiotic treatment, combination therapy showed no advantage over treatment with a single beta-lactam drug except for neutropenic patients (odds ratio [OR] for mortality, 0.5; 95% confidence interval [95% CI], 0.2 to 1.3) and patients with Pseudomonas aeruginosa bacteremia (OR, 0.7; 95% CI, 0.3 to 1.8). On multivariable logistic regression analysis including all risk factors for mortality, combination therapy had no advantage over therapy with a single beta-lactam drug. The mortality rate for patients treated with a single appropriate aminoglycoside was higher than that for patients given a beta-lactam drug in all strata except for patients with urinary tract infections. When the results of blood cultures were known, 1,878 patients were available for follow-up. Of these, 816 patients were given a single beta-lactam drug, 442 were given combination treatment, and 193 were given a single aminoglycoside. The mortality rates were 13, 15, and 23%, respectively (P = 0.0001). Both on stratified and on multivariable logistic regression analyses, combination treatment showed a benefit over treatment with a single beta-lactam drug only for neutropenic patients (OR, 0.2; 95% CI, 0.05 to 0.7). In summary, combination treatment showed no advantage over treatment with an appropriate beta-lactam drug in nonneutropenic patients with gram-negative bacteremia.