In-vitro activity of combined ampicillin and ciprofloxacin against high- level gentamicin-resistant strains of Enterococcus faecium: influence of the medium on susceptibility test results

Sir, The optimum therapy of patients with severe infections caused by high-level gentamicin-resistant (HLGR) enterococci has not yet been determined. In a previous study, the combination of ampicillin and ciprofloxacin was found to be bactericidal against such strains in vitro. However, the same combination failed to sterilize valvular vegetations in an experimental endocarditis model. In an attempt to explain this discrepancy, we evaluated the effects of different media, varying the Ca and Mg concentrations of the medium and the presence or absence of human serum in the medium on the activities of the ampicillin– ciprofloxacin combination against five HLGR, ampicillinand ciprofloxacin-resistant clinical isolates of Entero coccus faecium. The MICs of ampicillin and ciprofloxacin (both obtained from Sigma Chemical Co., St Louis, MO, USA) were determined by a microbroth dilution method. The media used in the study were as follows: Mueller–Hinton broth (MHB; Difco, Detroit, MI, USA), cation-adjusted Mueller–Hinton II broth (MHB II; BBL Becton Dickinson, Cockeysville, MD, USA), Brain Heart Infusion broth (BHI; Difco), MHB supplemented with 50% pooled human serum (HS-MHB) and MHB supplemented with Ca and Mg to give final concentrations of 50 mg/L and 25 mg/L respectively (Ca/Mg-MHB). Time–kill studies were performed with ampicillin alone, ciprofloxacin alone, both ampicillin and ciprofloxacin or no antibiotic (control) according to a method described previously; the concentrations of ampicillin and ciprofloxacin were 20 mg/L and 2 mg/L respectively. An overnight culture of each strain was diluted in each of the five media to give suspensions that initially contained 10 cfu/L. Bactericidal activity was defined as a reduction in the initial suspension of 3 log10 cfu/L after incubation at 37oC for 24 h. All of the experiments were performed at least twice and the results expressed as the means. The MICs of ampicillin and ciprofloxacin for the five enterococcal strains in MHB, MHB II, BHI and Ca/Mg-MHB were 32–64 mg/L and 2–4 mg/L respectively. The MICs of ampicillin in HS-MHB were markedly higher, i.e. 128–256 mg/L, while those of ciprofloxacin in this medium were only slightly higher, i.e. 4–8 mg/L. The results of the time–kill studies are shown in the Table. Ampicillin consistently failed to exhibit more than marginal inhibitory activity against any of the strains. Ciprofloxacin, on the other hand, exhibited bacteriostatic activity against all strains in MHB, MHB II and BHI, although its inhibitory activities in HS-MHB and Ca/Mg-MHB were minimal. In attempting to correlate bacterial survival in the presence of ciprofloxacin with the cation content of the medium (Table), we demonstrated a significant inverse relationship between ciprofloxacin inhibitory activity and Mg concentration with all strains (r 0.84–0.94; P 0.05) and Ca concentration with four of the five strains (r 0.86–0.94; P 0.05). These data confirm the earlier observation that divalent cations antagonize the in-vitro activities of quinolone antibiotics. The combination of ampicillin and ciprofloxacin was bactericidal against all strains in MHB and against one of the five strains in MHB II, although, for the remaining four strains, the effect was just below the criterion that defined bactericidal activity, i.e. 2.56–2.99 log10 cfu/L; synergy between the two agents was observed with all strains in both media. In contrast, the combination exhibited only bacteriostatic activity against all strains in BHI and HS-MHB. Finally, in Ca/Mg-MHB, it was bactericidal against two strains and exhibited activity that was just below the threshold that defined bactericidal activity (2.37–2.84 log10 cfu/L) for the remaining three strains; synergy was observed with all five isolates. The bactericidal activity of the combination in this medium was therefore similar to that in MHB, even though the activity of ciprofloxacin in the former was markedly lower than that in the latter. The failure of the antibiotic combination to exhibit bactericidal activity in enriched media (BHI and HS-