ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Apr. 1979, p. 592-596 Vol. 15, No. 4 0066-4804/79/04-0592/05$02.00/0 Effect of Time and Concentration Upon Interaction Between Gentamicin, Tobramycin, Netilmicin, or Amikacin and Carbenicillin or Ticarcillin LARRY K. PICKERING* AND PAM GEARHART Program in Infectious Diseases and Clinical Microbiology and Department ofPediatrics, The University of Texas Medical School, Houston, Texas 77030 Received for publication 18 January 1978 An aminoglycoside antibiotic and carbenicillin or ticarcillin are widely used in the treatment of patients with gram-negative bacillus infections. This study evaluated the effect of time upon in vitro interaction between mixtures of four aminoglycosides at two concentrations with carbenicillin or ticarcillin at four concentrations. By linear regression analysis, the inactivation of each aminogly- coside was shown to be directly proportional to the concentration of carbencillin (P < 0.001). Inactivation was significantly (P < 0.01) greater for gentamicin and tobramycin than for amikacin or netilmicin at all carbenicillin concentrations. At carbenicillin concentrations of 300 and 600 i.g/ml, significantly (P < 0.005) less inactivation of amikacin occurred when compared to netilmicin. Ticarcillin pro- duced a significant (P < 0.025) inactivation of gentamicin and tobramycin, with inactivation being directly proportional to ticarcillin concentration. No inactiva- tion ofamikacin or netilmicin activity occurred unless the ticarcillin concentration was 600 ,ug/ml. No significant change in aminoglycoside activity occurred when stored with ticarcillin or carbenicillin at concentrations ranging from 100 to 600 ug/ml at -70°C for 56 days. When an aminoglycoside and carbenicillin or ticarcillin are indicated in patients with renal failure, this study supports the use of ticarcillin with either amikacin or netilmicin. Aminoglycoside antibiotics in combination found that a 25 to 74% reduction in the half-life with carbenicillin or ticarcillin frequently are of gentamicin resulted from the concomitant used for treatment of patients with infections administration of therapeutic doses of carbeni- due to gram-negative bacilli (3, 8, 12; J. L. cillin and ticarcillin. Holt et al. (10) showed that Hoecker, L. K. Pickering, D. Groschel, S. Kohl, carbenicillin and ticarcillin inactivated genta- and J. van Eys, Cancer, in press). The combi- micin, tobramycin, sisomicin, and amikacin at a nation of carbenicillin or ticarcillin with an ami- concentration of 50:1 and that this inactivation noglycoside provides a broad spectrum of activ- was least in pooled human sera and greatest in ity against gram-negative bacilli and has been phosphate buffer at pH 7.4. This study was shown to be synergistic against Pseudomonas conducted to determine the effect of time and aeruginosa and various Enterobacteriaceae (1, concentration upon the inactivation of four ami- 2, 6, 11, 13, 17). Conversely, inactivation of ami- noglycoside antibiotics when exposed to various noglycoside antibiotics has been shown to be concentrations of ticarcillin or carbenicillin. caused by either ticarcillin or carbenicillin (7, 10, (This paper was presented at the 18th Inter- 15, 16, 18, 21). McLaughlin and Reeves (16) science Conference on Antimicrobial Agents demonstrated in vitro inactivation ofgentamicin and Chemotherapy, Atlanta, Georgia, October, by carbenicillin and suggested that some antag- 1978.) onistic effect might occur in vivo by a combina- tion that had been regarded as synergistic. Riff MATERIALS AND METHODS and Jackson (21) studied the half-life of genta- micin in four patients receiving hemodialysis Amikacin base (902 u.g/mg; Bristol Laboratories, Syracuse, N.Y.), gentamicin sulfate (586 tg/mg), and and found that when carbenicillin and gentami- netilmicin sulfate (583 ,ug/mg; Schering Corporation, cin were administered concomitantly in a dosage Bloomfield, N.J.), tobramycin sulfate (968 gLg/mg; Eli ratio of 80:1, the half-life of gentamicin was Lilly & Co., Indianapolis, Ind.), disodium ticarcillin markedly reduced. Davies and associates (7) (870 ug/mg; Beecham Laboratories, Bristol, Tenn.), evaluated eight patients with renal failure and and disodium carbenicillin (Roerig, New York, N.Y.) 592 VOL. 15, 1979 FOUR AMINOGLYCOSIDES 593 were dissolved in pooled human sera obtained from three donors to achieve final concentrations of 5 and 10 ug/ml for gentamicin, netilmicin, and tobramycin, a- and 10 and 20,ug/ml for amikacin. At the same time, 80' these concentrations ofeach aminoglycoside antibiotic I- were mixed separately with 100, 200, 300, and 600 pg/ 4 60- w ml concentrations of ticarcillin or carbenicilin in 0 pooled human sera. All samples were adjusted to a pH of 7.4. Ou 40'- Immediately after mixing, a portion of each speci- -l 0 men was assayed for specific aminoglycoside activity. Z 20- Each mixture then was divided and subjected to the i following: (i) incubation at 370C, with further portions 4 being assayed for aminoglycoside activity at 1 and 3 days; (ii) storage at -70°C, with portions being thawed 0 100 200 300 400 500 600 and assayed for residual aminoglycoside activity at 1, CARBENICILLIN CONC. (pg/ml) 3, 7, and 56 days. Results of residual aminoglycoside FIG. 1. Linear regression analysis of gentamicin activity for each sample are expressed as percent (Gent.), tobramycin (Tobr.), netilmicin (Neti.), and activity when compared to activity at zero time. All amikacin (Amik.) activity after exposure to various samples were assayed in duplicate. concentrations of carbenicillin at 37°C for 72 h (n = Concentrations of gentamicin and tobramycin were 16 for each aminoglycoside). determined by radioenzymatic assay, using an ade- nyltransferase enzyme (4, 22). Netilmicin and amika- cin were determined by radioenzymatic assay, using 0.001). A similar inactivation of tobramycin oc- an acetyltransferase enzyme (5). Each assay has a curred when 5 and 10 pg of tobramycin per ml standard curve linear from 1 to 100 iLg/ml and a within- were incubated with 100 pg of carbenicillin per and-between assay precision of 6%. Samples contain- ml (1:20 and 1:10, respectively) or when 5 and 10 ing tobramycin and netilmicin also were assayed for pug of tobramycin per ml were exposed to 300 pg activity by using microbiological and radioimmune of carbenicillin per ml (1:60 and 1:30, respec- assays as previously reported by us (4, 5). Bacillus tively). No change in netilmicin activity oc- globogii and Klebsiellapneumoniae were used as the curred when exposed to 100 ug of test organisms for tobramycin and netilmicin, respec- carbenicillin tively. Results were compared and correlated with the per ml; however, as the concentration of carben- radioenzymatic assay results. icillin increased, inactivation occurred, most no- Statistical analysis was performed by using a paired ticeably after 72 h of incubation. This inactiva- t test within and between groups and linear regression tion was significantly (P < 0.01) less than seen analysis to determine the effect of carbenicillin con- in gentamicin or tobramycin at all carbenicillin centrations (Fig. 1). Values expressed represent mean concentrations studied. A similar degree of in- ± 1 standard error of the mean of three to four activation occurred when 5 and 10 pg of netil- separate experiments. micin per ml were incubated with 300 pg of carbenicillin per ml (1:60 and 1:30, respectively). RESULTS Decreases in amikacin activity were significantly Incubation at 37°C. The percent activity of (P < 0.001) less at all times and concentrations the various aminoglycoside antibiotics after in- tested than changes that occurred in gentamicin cubation with carbenicillin at 37°C at concentra- and tobramycin activity. At clinically relevant tion ranging from 100 to 600 Ag/ml for periods concentrations, amikacin inactivation was not of 1 to 3 days are shown in Table 1. The activity different than netilmicin inactivation at carben- of gentamicin, tobramycin, netilmicin, and icillin concentrations of 100 and 200 ,ug/ml; how- amikacin showed no significant decrease with ever, at carbenicillin concentrations of 300 and time when carbenicillin was not present. When 600 ug/ml, significantly (P < 0.005) less inacti- either 5 or 10 ,ig of gentamicin per ml was vation of amikacin occurred when compared to incubated with 100 ,ug of carbenicillin per ml, a netilmicin. significant inactivation of gentamicin occurred Figure 1 depicts linear regression analysis of after 24 h (P < 0.02) and 72 h (P < 0.005). As gentamicin, tobramycin, netilmicin, and amika- the concentration of carbenicillin increased, the cin activity when exposed to various concentra- inactivation of gentamicin became more signifi- tions of carbenicillin at 370C for 72 h. There was cant, with the greatest inactivation occurring at a direct relationship between the concentration carbenicillin concentrations of600pg/ml. Tobra- of carbenicillin and the inactivation of each ami- mycin and carbenicillin interaction results were noglycoside. The higher the concentration of similar to those of gentamicin and carbenicillin, carbenicillin in the reaction mixture, the greater with tobramycin inactivation being directly pro- the inactivation of gentamicin, tobramycin, portional to carbenicillin concentration (P < netilmicin, and amikacin (P < 0.001). 594 PICKERING AND GEARHART ANTIMICROB. AGENTS CHEMOTHER. Table 2 shows the percent activity of genta- micin, tobramycin, netilmicin, and amikacin when incubated with ticarcillin at concentra- - -