Antibiotic combinations for serious infections caused by carbapenem-resistant Acinetobacter baumannii in a mouse pneumonia model

Objectives: Successful therapy of carbapenem-resistant Acinetobacter baumannii strains has been reported with colistin, but recently we argued against its use as monotherapy because of the poor results obtained in a mouse pneumonia model. Our aim was to identify antibiotic combinations that were valid therapeutic alternatives in the same model. Methods: We used two carbapenem-resistant A. baumannii strains (D and E; MICs of imipenem, 8 and 512 mg/L, respectively). MICs of tobramycin, rifampicin and colistin for both strains were 8, 8 and 0.5 mg/L, respectively. Results: In infections caused by strain D, lung bacterial counts (log(10) cfu/g, mean +/- s.d.) were: controls (10.86+/-0.25), imipenem (5.99+/-0.59, P < 0.05 versus controls), and colistin (10.43 +/- 1.09); imipenem + tobramycin was the most active combination (5.46+/-0.62, P < 0.05 versus controls). In infections caused by strain E, results were: controls (10.82+/-0.33), rifampicin (5.62+/-0.26, P < 0.05 versus controls), colistin (8.38+/-1.22, P < 0.05 versus controls), and imipenem (11.01+/-0.2); rifampicin + imipenem (3.79+/-0.99) and rifampicin + tobramycin (3.96+/-0.30) were the most active combinations (P < 0.05); results with rifampicin + colistin (5.59+/-1.17) were similar to those with rifampicin alone. Conclusions: Our data indicate that imipenem can still be the best alternative for carbapenem-resistant A. baumannii infections with moderate levels of imipenem resistance, preferably combined with aminoglycosides. For strains highly resistant to imipenem, a combination of rifampicin with imipenem, tobramycin or colistin may be useful, if resistance to rifampicin is only moderate.