ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Jan. 1996, p. 173–178 0066-4804/96/$04.0010 Copyright q 1996, American Society for Microbiology
Vol. 40, No. 1
Treatment of Experimental Endocarditis Due to Enterococcus faecalis Using Once-Daily Dosing Regimen of Gentamicin plus Simulated Profiles of Ampicillin in Human Serum ` ,1* PERE JOAN CARDONA,1 BENITO ALMIRANTE,1 JOSEP ANTO ´ N CAPDEVILA,1 JOAN GAVALDA MONTSERRAT LAGUARDA,1 LEONOR POU,2 ERNESTO CRESPO,3 CARLES PIGRAU,1 AND ALBERT PAHISSA1 Infectious Diseases Research Laboratory, Infectious Diseases Division,1 and Biochemistry2 and Microbiology3 Services, Hospital Universitari Vall d’Hebron, Universitat Auto `noma de Barcelona, Barcelona, Spain Received 18 July 1995/Returned for modification 1 September 1995/Accepted 3 November 1995
We compared the efficacy of ampicillin, both alone and in combination with gentamicin given once a day (q.d.) or three times a day (t.i.d.), in the treatment of experimental enterococcal endocarditis. Ampicillin was administered by using humanlike pharmacokinetics that simulated the profiles of this drug in human serum. An open one-compartment mathematical model developed in this study was used to estimate the decreasing doses administered with a computer-controlled infusion pump that simulated in rabbits the human serum pharmacokinetics after intravenous administration of 2 g of ampicillin every 4 h. Animals with catheterinduced endocarditis were infected intravenously with 108 CFU of Enterococcus faecalis J4 (MICs and MBCs of ampicillin and gentamicin, 2 and 128 and 16 and 64 mg/ml, respectively) and were treated for 3 days with ampicillin alone or in combination with gentamicin at 2 mg/kg of body weight subcutaneously t.i.d. or at 6 mg/kg subcutaneously q.d. The serum ampicillin levels and pharmacokinetic parameters of the humanlike pharmacokinetics of ampicillin in rabbits were similar to those found in humans treated with 2 g of ampicillin intravenously. The results of therapy for experimental endocarditis caused by E. faecalis J4 showed that the residual bacterial concentration in aortic valve vegetation was significantly lower in the animals treated with combinations of ampicillin plus gentamicin given q.d. or t.i.d. than in those treated with ampicillin alone (P < 0.01). The dosing interval of gentamicin did not significantly affect (q.d. versus t.i.d.; P 5 0.673) the therapeutic efficacy of the combination of ampicillin plus gentamicin. ematical model that enabled us to determine the doses to be administered to rabbits by a computer-controlled infusion pump system to simulate the human kinetics of an antimicrobial agent. Thus, the elimination kinetics of ampicillin were examined in healthy rabbits, and thereafter ampicillin was administered to animals in a way that simulated the profile of ampicillin in human serum after an intravenous bolus dose of 2 g. The other purpose of our study was to investigate the therapeutic outcome of the combination of ampicillin plus gentamicin given once a day, using humanlike pharmacokinetics of ampicillin, in the treatment of experimental endocarditis due to Enterococcus faecalis. (This work has previously been presented in part [17a]).
Four to 6 weeks of penicillin or ampicillin plus an aminoglycoside is effective treatment for enterococcal endocarditis (4). Because of the potentially toxic effects of prolonged aminoglycoside therapy, especially in elderly patients or those with impaired renal function, less toxic but equally effective therapeutical strategies are being investigated. Recent experimental (3, 23, 24, 28, 36) and clinical evidence (25, 31, 33–35) concerning gram-negative rod infections has suggested that aminoglycoside administration in one total daily dose has the same therapeutic effectiveness as a regimen of divided doses, with comparable or reduced toxicity. Moreover, clinical (14) and experimental studies (5, 6, 8, 18, 32) have shown that increasing the dosing interval of aminoglycosides does not significantly affect their efficacy in the treatment of penicillin-susceptible or -resistant viridans group streptococcal endocarditis. This therapeutic approach has not been extensively studied in the treatment of enterococcal endocarditis (11) and merits further investigation. Studies of antimicrobial efficacy in experimental models of infection provide an important basis for clinical investigative studies in humans, but antibiotic pharmacokinetics may differ greatly between humans and animals because of the faster elimination rates of drugs in animals. In order to surmount this problem, several authors have used animal models of humanlike pharmacokinetics (5, 12, 13, 19– 21, 26, 36–38). One purpose of this study was to develop a new and amenable open one-compartment pharmacokinetic math-
MATERIALS AND METHODS In vitro studies. (i) Test strain. E. faecalis J4 was isolated from the blood of a patient with endocarditis. This strain, which was susceptible to ampicillin and moderately resistant to gentamicin, was used to study the in vitro and in vivo activities of antimicrobial agents. Working stock cultures were kept frozen at 2808C in skim milk (Difco Laboratories, Detroit, Mich.). Before each experiment, one aliquot was thawed and subcultured onto 5% sheep blood Columbia ´ toile, France). agar plates (bioMe´rieux, Marcy l’E (ii) Media and antibiotics. Mu ¨eller-Hinton broth and 5% sheep blood Mu ¨eller-Hinton agar plates (bioMe´rieux) were used. Drugs were supplied by the following manufacturers: ampicillin by Antibio ´ticos SA (Madrid, Spain) and gentamicin by Schering-Plough (Madrid, Spain). Fresh ampicillin solutions were prepared on the day they were used. Stock solutions of gentamicin were prepared and stored at 2208C. (iii) In vitro antibiotic susceptibility tests. MICs and MBCs were determined in triplicate by the microdilution method (2). Inocula were prepared from logarithmic-phase cultures to yield a final concentration of 105 CFU/ml and inoculated into serial twofold dilutions of ampicillin or gentamicin in Mu ¨eller-Hinton broth supplemented with 3% lysed sheep blood (bioMe´rieux). Staphylococcus aureus ATCC 25923 was used as the control test strain. Subcultures were made
* Corresponding author. Mailing address: Infectious Diseases Division, Hospital General Universitari Vall d’Hebron, Avda. Vall d’Hebron, 119-129, 08035 Barcelona, Spain. Phone: 34.3.2463971. Fax: 34.3.4282637. Electronic mail address:
[email protected]. 173
174
` ET AL. GAVALDA
to confirm purity and to quantify the inoculum size. Microtiter plates were incubated at 378C in room air. The MIC was defined as the lowest concentration of the antimicrobial agent that did not allow visible growth after 24 h of incubation. To determine the MBC, aliquots of 0.025 ml each were taken from the control well, from the first well with visible growth, and from each dilution of antibiotic that did not exhibit turbidity after 24 h and were plated onto 5% sheep blood Mu ¨eller-Hinton agar plates. After 48 h of incubation at 378C in room air, viable bacteria were counted. The MBC was defined as the lowest concentration of the antimicrobial agent that killed at least 99.9% of the original inoculum (2). (iv) Studies of combined antimicrobial activity. Time-kill curves were used to test the bactericidal activities of ampicillin and gentamicin alone and in combination against E. faecalis J4. Killing curves were determined by using Mu ¨ellerHinton broth supplemented with 3% lysed sheep blood and an inoculum of 106 CFU/ml while the organism was in the logarithmic phase of growth. Antibiotics, both alone and in combination, were added at final concentrations of 5 mg of ampicillin per ml and 0.25, 0.5, or 1 mg of gentamicin per ml. After 0, 4, and 24 h of incubation at 378C in room air, serial dilutions of 0.05-ml samples (to 1026) were subcultured onto 5% sheep blood Columbia agar plates. After incubation for 48 h at 378C in room air, CFU were counted and plotted against time. Each experiment was run twice in duplicate and included controls of growth without antibiotic. Synergism was defined as at least a 100-fold increase in killing at 24 h with the combination of ampicillin and gentamicin in comparison with the most active single drug (10). Serum ampicillin pharmacokinetic studies. Ampicillin was administered by using a system designed to reproduce human serum pharmacokinetics in rabbits in order to mimic the profile of ampicillin in human serum after an intravenous (i.v.) infusion of 2 g of ampicillin. A computer-controlled infusion pump system that delivered decreasing quantities of drug infusion pump (Alice King) was employed. The computer software was written by us. This approach involved three steps. (i) The ampicillin pharmacokinetic parameters in rabbits were estimated. An open one-compartment model was used to estimate the elimination rate constant (kel) and the volume of distribution (V) after an i.v. injection of ampicillin in six healthy New Zealand White rabbits. Although the antimicrobial agents do not have compartment models, for practical reasons we have used an open one-compartment model to fit the data. To determine the concentrations of ampicillin in serum samples, blood was drawn from a carotid catheter at 10, 20, 30, 45, 60, 90, 120, and 180 min after a single i.v. injection of 100 mg of ampicillin/kg. This study was done with six healthy rabbits. Blood samples were immediately placed into tubes and centrifuged (10 min, 1,500 3 g). The upper phase was placed at 48C and assayed the same day. Ampicillin concentrations were determined by the disk-plate bioassay method (1) with Micrococcus luteus ATCC 9341 as the bioassay microorganism and antibiotic medium 5 (Difco Laboratories) as the growth medium. Standard curves were determined with solutions of ampicillin (1, 2, 3, 4, and 5 mg/ml) in pooled rabbit serum. The concentrations in serum samples were derived from the standard curve. Serum samples from rabbits were diluted with pooled rabbit serum so that their concentrations would be within the range of the standard curve. Standard samples were assayed in quintuplicate, and serum samples were assayed in triplicate. The results were expressed in micrograms per milliliter of blood. The linearity of the standard curve was 0.995. The sensitivity of this assay was about 1 mg/ml of sample, and the coefficients of between- and within-day variation for replicates (n 5 7) at 1 and 40 mg/ml were less than 5%. The kel and the concentration in serum at zero time in rabbits (kelr and C0r, respectively) were determined from a linear regression analysis of the concentration-time curve on the basis of an open one-compartment model. The Vr (in liters per kilogram) was calculated as the dose given i.v. (in milligrams per kilogram)/C0r (in milligrams per liter). (ii) A mathematical model was applied to obtain the required infusion doses to simulate human kinetics in animals. The development of this mathematical model is shown in Appendix. (iii) In vivo experimental pharmacokinetic studies were done to simulate in rabbits the pharmacokinetic profile of ampicillin in humans. Briefly, two polyethylene catheters (inside diameter, 0.81 mm; outside diameter, 1.27 mm; Portex SA, Hythe, Kent, England), one through the carotid artery (sampling) and the other into the cava vein through the jugular vein (infusion), were inserted. Both lines were tunnelled subcutaneously and brought to the intercapsular region. The external portion of the jugular catheter was connected to a flowthrough swivel and then to a computer-controlled infusion pump system. The pump system was set up to deliver previously calculated flow rates of i.v. infusion to simulate the human kinetics of 2 g of ampicillin given i.v. This study was done with seven healthy rabbits. To determine serum ampicillin concentrations, 2 ml of blood was sampled at 5, 15, 30, 60, 90, 120, 180, and 240 min after the start of infusion through the carotid catheter, immediately placed into a tube, and centrifuged (10 min, 1,500 3 g). Serum ampicillin concentrations were assayed on the same day by the microbiological bioassay described above. Different pharmacokinetic parameters were estimated on the basis of an open one-compartment model to compare the pharmacokinetics of ampicillin in rabbits, the human-adapted model, and humans. The half-life (t1/2) at b phase of ampicillin in rabbits with humanlike pharmacokinetics was calculated as ln 2/kel. kel was determined as the slope obtained from a linear regression analysis of the concentration-time curve on the basis of an open one-compartment model.
ANTIMICROB. AGENTS CHEMOTHER. Thereafter, the area under the concentration-time curve from 0 h to infinity (AUC0–`) was calculated as C0/kel, and the time over MIC was calculated as (ln C0 2 ln MIC)/kel. The pharmacokinetic parameters of ampicillin in humans and rabbits were calculated as described above with a C0 of 80 mg/ml. Serum gentamicin pharmacokinetic studies. Serum gentamicin levels were determined in five healthy New Zealand White rabbits after a single subcutaneous injection of 2 or 6 mg/kg. Two milliliters of blood was sampled at 15, 30, 60, 90, 120, 180, 240, 360, and 480 min after injection through a carotid catheter, immediately placed into a glass tube, and centrifuged (10 min, 1,500 3 g). Serum samples were stored at 2708C, and all samples were assayed on the same day. The concentrations of gentamicin in serum samples were measured by fluorescence polarization (TDx; Abbott Diagnostics, Irving, Tex.) according to the manufacturer’s directions. The within-day precision and between-day precision of this assay were 2.5 and 4.6%, respectively. The lower limit of detection was 0.25 mg of gentamicin per ml of serum. Establishment of endocarditis and installation of the infusion pump system. Experimental aortic valve infective endocarditis was induced in New Zealand White rabbits (approximately 2 to 2.1 kg each) by the method of Garrison and Freedman (17, 31) as modified by Durack and Beeson (9). The induction of nonbacterial thrombotic endocarditis was done as previously described (18). Twenty-four hours after the placement of intracardiac catheters, rabbits were inoculated via the marginal ear vein with 1 ml of saline containing 108 CFU of E. faecalis J4. The day before each experiment, a control stock culture of E. faecalis J4 was thawed and subcultured onto a 5% sheep blood Columbia agar plate. Inocula were set up by overnight subculture of three to four colonies in brain heart broth (bioMe´rieux). On the same day, a catheter (inside diameter, 0.81 mm; outside diameter, 1.27 mm; Portex SA) was placed into the superior cava vein through the jugular vein by the same technique described above to administer ampicillin treatment. The infusion pump system was set up to deliver 2 ml of 0.9% saline per h to keep the catheter open until the beginning of ampicillin administration. One milliliter of blood was obtained 24 h after infection and just before the initiation of antimicrobial therapy to confirm the presence of endocarditis. The blood specimen was mixed with 10 ml of molten tryptic soy agar (bioMe´rieux). Plates were incubated for 48 h at 378C in room air, and the presence of E. faecalis was interpreted as indicative of infective endocarditis. Treatment groups and estimation of therapeutic efficacy. Antimicrobial therapy was initiated 24 h after infection and continued for 3 days. Ampicillin was administered i.v. with the computer-controlled infusion pump system, which delivered decreasing quantities of this drug to reproduce ampicillin profiles in human serum (humanlike pharmacokinetics). Rabbits were randomized to the following treatment groups: (i) control without treatment; (ii) ampicillin (humanlike pharmacokinetics), 2 g every 4 h i.v.; (iii) ampicillin (humanlike pharmacokinetics), 2 g every 4 h i.v., plus gentamicin, 2 mg/kg every 8 h; and (iv) ampicillin (humanlike pharmacokinetics), 2 g every 4 h i.v., plus gentamicin, 6 mg/kg every 24 h. Gentamicin was administered subcutaneously 0.5 h before the peak concentration of ampicillin. Ampicillin was dissolved in 0.9% saline, and the pump system was maintained at 48C during the treatment period to counteract ampicillin degradation. Antimicrobial solutions were changed every 24 h. After 3 days of treatment, animals were sacrificed 6 h after the ampicillin infusion was ended with a lethal i.v. injection of sodium pentothal. The chest was opened, the heart was excised and opened, and aortic valve vegetations were removed aseptically. Only vegetations adherent to the aortic valve were considered for further study. Inclusion in this study required verification of proper placement of the catheter across the aortic valve at autopsy and a positive blood culture, which was obtained prior to the onset of treatment. The animals in the control group were sacrificed 24 h after the induction of infection. Vegetations were rinsed with saline, weighed, and homogenized in 2 ml of tryptic soy broth in a tissue homogenizer (stomacher 80r). Homogenates were serially diluted and quantitatively cultured onto 5% sheep blood Columbia agar plates. Plates were incubated for 48 h at 378C in room air. The results were expressed as the log10 CFU of E. faecalis per gram of vegetation. Analysis of results. The results were expressed as the mean 6 standard deviation (SD), median, and range of log10 CFU of E. faecalis per gram of vegetation. Differences in log10 CFU of enterococci per gram were analyzed statistically by using the Kruskall-Wallis test and Wilcoxon-Mann-Whitney U rank sum analysis, and the results were corrected for multiple comparisons. A P value of less than 0.05 was considered significant.
RESULTS In vitro studies. The MIC and MBC of ampicillin were 2 and 128 mg/ml, while the MIC and MBC of gentamicin were 16 and 64 mg/ml, respectively. The in vitro killing of E. faecalis J4 by ampicillin and gentamicin, both alone and in combination, is shown in Fig. 1. The time-kill curve shows that gentamicin alone had no antimicrobial effect. Ampicillin alone was only bacteriostatic (the viable counts in culture did not significantly decrease after 24 h of incubation). In contrast, combinations of ampicillin and 0.5 or 1 mg of gentamicin per ml were rapidly
VOL. 40, 1996
GENTAMICIN AND EXPERIMENTAL ENTEROCOCCAL ENDOCARDITIS
175
FIG. 1. Killing curves obtained with ampicillin-gentamicin against E. faecalis J4. Ampicillin (A) was used at 5 mg/ml, and gentamicin (G) was used at 0.25, 0.5, and 1 mg/ml.
bactericidal and significantly killed strain J4, with a reduction in bacterial numbers of approximately 3 log10 units during the first 4 h; thus, these combinations were rapidly synergistic against strain J4. The combination of ampicillin and 0.25 mg of gentamicin per ml was not synergistic against this strain. Pharmacokinetic studies. The rabbit pharmacokinetic data we used in the mathematical model (determined in healthy rabbits that had received one intravenous bolus dose of 100 mg/kg) are shown in Table 1. The computer-controlled infusion pump system enabled us to produce serum ampicillin kinetics in rabbits similar to those found in humans after an i.v. 2-g injection (Fig. 2A). The pharmacokinetic parameters obtained from the human-adapted model were similar to those of i.v. ampicillin in humans (Table 1). Data from this model suggest that ampicillin trough concentrations would be above the MIC for the entire dosing interval, as observed in humans. The mean 6 SD of serum gentamicin levels after one subcutaneous injection of gentamicin at the various dose levels are shown in Fig. 2B. The peak serum gentamicin concentration, obtained 30 min after subcutaneous injection of 2 mg/kg, was
TABLE 1. Comparison of pharmacokinetic parametersa Value (mean 6 SD) Parameter (unit)
t1/2 (h) kel (h21) AUC0–` (mg z h/ml) Time over MIC (h) a
Rabbit (n 5 6)b
Human 2 g i.v.c
Rabbit humanlike (n 5 7)d
0.3 6 0.03 2.4 6 0.29 38.19 1.79
1.1 0.63 127.8 5.89
0.99 6 0.08 0.71 6 0.05 116.7 6 31.83 5.22 6 0.3
The parmacokinetic parameters were estimated on the basis of an open one-compartment model. b t1/2 and kel data are for healthy rabbits treated with 100 mg/kg i.v. The AUC and time over MIC data were calculated from an ideal profile obtained with a C0r of 80 mg/ml and the kel of ampicillin in rabbits. c kel and C0 data were obtained from reference 29. Data are for 2 g of ampicillin given i.v. to humans. d Data are for humanlike pharmacokinetics of 2 g of ampicillin given i.v. to rabbits.
FIG. 2. Results of pharmacokinetic studies of rabbits using humanlike pharmacokinetics of 2 g of ampicillin i.v. (A) or gentamicin at 2 or 6 mg/kg given subcutaneously (sbc) (B).
4.25 6 0.29 mg/ml, similar to that recommended in humans for the treatment of enterococcal endocarditis. The t1/2 of gentamicin given subcutaneously was 1.45 6 0.16 h (at 2 mg/kg) or 1.69 6 0.73 h (at 6 mg/kg), similar to the t1/2 of gentamicin given i.v. to humans (2 to 3 h). Treatment of established endocarditis. Fifty-six animals underwent the catheterization procedure. Ten were excluded from analysis because of technical problems with the computer-controlled infusion pump system. Two animals expired from
176
` ET AL. GAVALDA
ANTIMICROB. AGENTS CHEMOTHER.
TABLE 2. Treatment of experimental endocarditis caused by E. faecalis J4 with a humanlike profile of ampicillin alone or in combination with gentamicin Treatment groupa (n)
Control without treatment (9) A at humanlike 2 g i.v./4 h (10) A at humanlike 2 g i.v./4 h 1 G at 2 mg/kg/8 h s.c. (8) A at humanlike 2 g i.v./4 h 1 G at 6 mg/kg/24 h s.c. (9) a b
Log10 CFU/g of vegetation Mean 6 SD
Median (range)b
11.23 6 0.6 7.7 6 0.54 5.95 6 0.49 6.11 6 0.75
11.6 (10.39–11.85) 7.5 (7.13–8.84)* 6 (5.26–6.6)*† 6.1 (4.95–7.42)*†‡
A, ampicillin; G, gentamicin; s.c., subcutaneously. p, P , 0.001 versus control; †, P , 0.01 versus ampicillin alone; ‡, P 5 0.673 versus ampicillin plus gentamicin at 2 mg/kg/8 h s.c.
infection prior to randomization, and six died before the third day of treatment (two in each group). Two animals with negative blood cultures were excluded. A final total of 36 animals were evaluated. The results of the different therapeutic regimens are shown in Table 2. All nine control rabbits had infected vegetations with high mean bacterial counts per gram of vegetation. After 3 days of treatment, the bacterial counts in the vegetations of treated animals were reduced compared with those of the control group for all drug regimens (P , 0.001). Comparisons between treated groups revealed that combinations of gentamicin and ampicillin were significantly more effective than was ampicillin alone in reducing the number of bacteria in vegetations (P , 0.01). Ampicillin, administered at the simulated human profile of 2 g i.v. every 4 h, and gentamicin, 6 mg/kg once a day, disclosed an activity similar to that of the gold standard, ampicillin combined with gentamicin at 2 mg/kg three times daily (P 5 0.673). None of the animals had sterile vegetations. DISCUSSION The usual approach to therapy in the rabbit model of endocarditis involves the administration of antimicrobial agents in doses that are calculated to achieve peak levels in serum within the range of values obtained in humans. In the experiments reported here, we used a mathematical model that determines the doses to be given to animals by a computer-controlled infusion pump system to obtain profiles of ampicillin in rabbit serum similar to those observed in human serum after i.v. administration of 2 g of ampicillin. This model can be used for other drugs with an open one-compartment pharmacokinetic model. The other purpose of this study was to study the efficacy of once-daily gentamicin combined with ampicillin, administered at humanlike pharmacokinetics, for the treatment of experimental enterococcal endocarditis. Our results show that a single daily dose of gentamicin combined with ampicillin was as effective as a divided regimen of the same dose of gentamicin plus ampicillin. Studies of the activities of antimicrobial agents in experimental animal models are a necessary basis for investigative studies in humans (28). Interspecies pharmacokinetic variations are a relevant factor to consider in these studies since pharmacokinetics affects antimicrobial in vivo activity (19, 21). Usually, drugs are administered to reach peak concentrations in serum similar to those in human serum; however, the faster elimination rates of drugs in animals are often not contemplated. To overcome this problem, animal models of humanadapted pharmacokinetics (5, 12, 13, 19–21, 26, 37–39) in which the drug is administered to produce serum kinetics comparable to that in humans have been described. An overview of these studies clarifies certain issues. Questions about therapeu-
tic efficacy that could not be answered previously have been elucidated (5, 12); moreover, animal kinetics may lead to overestimation (for aminoglycosides) or underestimation (b-lactams) of antimicrobial activity compared with results from human-adapted models (19, 21). Thus, the use of humanlike animal pharmacokinetic models would have some significance in the interpretation of animal studies that evaluate the efficacies of different types of antimicrobial agents (e.g., b-lactams versus glycopeptides) or antimicrobial combinations. Our results show that the model we have developed is simple and adaptable for simulating the human kinetics of antimicrobial agents with an open one-compartment model. Since gentamicin is an antibiotic with an open three-compartment model, our humanlike kinetic model could not be applied to this drug. We chose administration by the subcutaneous route because it induces a serum elimination profile in rabbits with an estimated t1/2 at b phase from 1.5 to 1.7 h, close to that found in humans (2 to 3 h). The advantage of once-daily administration of aminoglycosides for the treatment of endocarditis is presently under discussion (5, 6, 8, 11, 18, 33). Tobramycin once a day and penicillin were found to be effective in treating experimental endocarditis due to nutritionally variant viridans group streptococci (33). It has been demonstrated that increasing the gentamicin interval to once a day does not affect the efficacy of penicillin combined with gentamicin in studies using penicillinsusceptible (6, 8, 18), -tolerant (18), and -resistant (18) viridans group streptococcal endocarditis models. Finally, Blatter et al. (5) have shown that ceftriaxone combined with netilmicin given once a day is more effective than is either agent alone in the treatment of viridans group streptococcal experimental endocarditis. On the basis of these results, a successful clinical trial using this combination for 2 weeks to treat patients with streptococcal endocarditis was conducted (14). In contrast to our results, Fantin and Carbon (11) reported that penicillin plus netilmicin at a dose that reached a peak level in serum of 9.8 mg/ml given three times daily was more effective than was the regimen of penicillin plus the same total daily dose of netilmicin given once a day in the treatment of enterococcal experimental endocarditis. In their study, a decrease in the dose of netilmicin (peak concentration in serum of 5.6 mg/ml) did not achieve in vivo synergism with penicillin. The most probable explanation for the discrepancies between our results and those of Fantin and Carbon is that the experimental variables in the two studies are distinct, a fact that makes comparison difficult. First, the b-lactams used (penicillin versus ampicillin), although similar, may possess different activities against E. faecalis (27). Second, our experimental kinetic model simulated the human kinetics of ampicillin after one i.v. 2-g dose every 4 h. A relationship between the shape of the AUC and b-lactam antimicrobial activity has been shown (19, 21); it may also affect in vivo efficacy when they are com-
VOL. 40, 1996
GENTAMICIN AND EXPERIMENTAL ENTEROCOCCAL ENDOCARDITIS
bined with aminoglycosides. Finally, treatment was started at different times after infection, a fact that can influence the outcome of experimental endocarditis treatment (7). The efficacy that we observed for the regimen of ampicillin plus gentamicin once a day in the treatment of enterococcal experimental endocarditis might be due to a postantibiotic effect (PAE) of this combination. Previous reports have shown synergistic increases in in vitro PAEs for combinations of various b-lactams and aminoglycosides against E. faecalis (15, 16, 22). A study by Hessen et al., however, that used the enterococcal endocarditis animal model failed to confirm an in vivo PAE after treatment with penicillin combined with gentamicin (22). To our knowledge, no other studies exclude the in vivo PAE against enterococci. In the study of Hessen et al., the peak concentration in vegetation was reported to be 2.1 mg/ml at 1 h; since increasingly long PAEs were seen with increasing concentrations of gentamicin in vitro, an increased dose of gentamicin in vivo could conceivably produce a detectable in vivo PAE (22). This event may have occurred in our experiments. Gentamicin given once a day produces a peak concentration in serum of 13.6 mg/ml, and peak concentrations in vegetations would also be higher. Another variable is that the use of humanlike kinetics produced serum ampicillin levels above the MIC for enterococci throughout the treatment interval; thus, the pharmacodynamics in vegetation may be considered different from that in the study of Hessen et al. (22). It is difficult to ensure that the efficacy of ampicillin combined with gentamicin once a day may not be explained in part by an in vivo PAE of this combination against enterococci. In conclusion, the results presented here show that in the treatment of E. faecalis-induced experimental endocarditis, one daily dose of gentamicin (to reach peak concentrations in serum of between 13 and 14 mg/ml) combined with ampicillin (given at humanlike kinetics of 2 g every 4 h) was as effective as the gold standard, recommended enterococcal endocarditis treatment in humans, ampicillin plus gentamicin given at a therapeutic schedule to reach peak concentrations in serum of 3 mg/ml three times a day. Discrepancies with other studies may be explained by differences in experimental variables. More data from experimental studies are necessary before this therapeutic schedule may be used for humans.
APPENDIX The development of this mathematical model is discussed herein. The aim of the mathematical model was to find the doses required to obtain the desired humanlike pharmacokinetics in these animals. The system we used to imitate human kinetics in rabbits is based on the administration of decreasing quantities of drug at previously established times (Tx [T1 to Tn]) to counteract the faster elimination rates in animals (Fig. A1). (i) At the final limit point of each time period (Tx), the serum drug level is higher in human kinetics (Chx) than in rabbit kinetics (Crx). We determine the concentration of the drug (Cdx [Cd1 to Cdn]) necessary to counterbalance this difference by subtracting the desired concentration (Chx [Ch1 to Chn]) from the estimated level in animal serum (Crx [Cr1 to Crn]) according to the following mathematical formulas, where x is the number of the period: for T1, Cd1 5 Ch1 2 Cr1 for T2, Cd2 5 Ch2 2 Cr2 where Ch1 5 C0 z e2kelh z t1, Cr1 5 C0 z e2kelr z T1, Ch2 5 C0 z e2kelh z t2, and 2kelr z T2 Cr2 5 Ch1 z e . Thus, the general formula for Tx 5 1 to infinity is Cdx 5 (C0 z e2kelh z tx) 2 (Chx-1 z e2kelr z Tx)
177
FIG. A1. Humanlike pharmacokinetics. Graphic expression of the mathematical model for drugs with an open one-compartment model. See Appendix for details.
where Tx is the time value between periods (in hours), tx is the time in human kinetics (Chx; in hours), kelh is the human elimination rate constant (h21), and kelr is the rabbit elimination rate constant (h21). (ii) The amount of drug (Qx; in milligrams per hour) that must be given by continuous i.v. infusion during the entire time period (Tx) to attain the desired concentration (Chx) at the end of the Tx may be determined by the following mathematical fraction: Qx 5 Cdx z kelr z Vr z Wr/(1 2 e2kelr z Tx) where Qx is the quantity of drug to be administered during Tx (in milligrams per hour), Vr is the volume of distribution of the drug in rabbits (in liters per kilogram), and Wr is animal weight (in kilograms). (iii) The infusion rate (Vx; in milliliters per hour) to use during Tx in order to administer Qx depends on the concentration of the drug solution used (S; in milligrams per milliliter), as follows: Vx 5 Qx/S or Cdx z kelr z Vr z Wr/[(1 2 e2kelr z Tx) z S] Thus, the Vx of an antimicrobial solution with a concentration of S to use during Tx to counterbalance the higher kel of rabbits compared with that of humans with the aim of imitating human elimination kinetics in an one-compartment pharmacokinetic drug model may be calculated with the following final formula: Vx 5 {[(C0 z e 2 kelh z tx) 2 (Chx-1 z e2kelr z Tx)] z kelr z Vr z Wr}/[(1 2 e 2 kelr z Tx) z S] The first i.v. dose is determined by the following formula: Vx 5 C0 z kelr z Vr z Wr/[(1 2 e2kelr z Tx) z S] where C0 is the peak concentration of the studied drug in serum. The human pharmacokinetic parameters (kelh and C0) after one i.v. dose of 2 g of ampicillin were obtained from the literature (29) and are shown in Table 1. ACKNOWLEDGMENTS We thank George M. Eliopoulos for providing helpful discussions and comments on the manuscript, Juan Molina for technical assistance
178
` ET AL. GAVALDA
ANTIMICROB. AGENTS CHEMOTHER.
in writing the computer program, and Celine Cavallo for English language assistance. This work was supported in part by Direccio ´n General de Investigacio ´n, Ciencia y Tecnologı´a grant no. PM91-0052.
19.
REFERENCES
20.
1. Anhalt, J. P. 1985. Antimicrobial assays, p. 691–729. In J. A. Washington II (ed.), Laboratory procedures in clinical microbiology. Springer-Verlag, New York. 2. Anhalt, J. P., and J. A. Washington II. 1985. Bactericidal tests, p. 731–745. In J. A. Washington II (ed.), Laboratory procedures in clinical microbiology. Springer-Verlag, New York. 3. Bamonte, F., S. Dionisotti, M. Gamba, E. Ongini, A. Arpini, and G. Melone. 1990. Relation of dosing regimen to aminoglycoside ototoxicity: evaluation of auditory damage in the guinea pig. Chemotherapy (Basel) 36:41–50. 4. Bisno, A. L., W. E. Dismukes, D. T. Durack, E. L. Kaplan, A. W. Karchmer, D. Kaye, S. H. Rahimtoola, M. A. Sande, J. P. Sanford, C. Watanakunakorn, and W. R. Wilson. 1989. Antimicrobial treatment of infective endocarditis due to viridans streptococci, enterococci, and staphylococci. JAMA 261: 1471–1477. 5. Blatter, M., U. Fluckiger, J. Entenza, M. P. Glauser, and P. Francioli. 1993. Simulated human serum profiles of one daily dose of ceftriaxone plus netilmicin in treatment of experimental streptococcal endocarditis. Antimicrob. Agents Chemother. 37:1971–1976. 6. Chambers, H. F., S. Kennedy, M. Fournier, M. S. Rouse, and W. R. Wilson. 1990. Ceftriaxone (C) vs. procaine penicillin (P) alone and with gentamicin (G) in viridans streptococcal aortic valve endocarditis in rabbits, abstr. 701, p. 203. In Program and abstracts of the 30th Interscience Conference on Antimicrobial Agents and Chemotherapy. American Society for Microbiology, Washington, D.C. 7. Cre´mieux, A. C., A. Saleh-Mghir, J. M. Vallois, M. Muffat-Jolly, C. Devine, J. J. Pocidalo, and C. Carbon. 1993. Influence of the pre-treatment duration of infection on the efficacies of various antibiotic regimens in experimental streptococcal endocarditis. J. Antimicrob. Chemother. 32:843–852. 8. Dorscher, C. W., B. M. Tallan, M. S. Rouse, J. M. Steckelberg, H. C. Chambers, and W. R. Wilson. 1990. Effect of gentamicin dosing interval on ceftriaxone or penicillin treatment (Rx) of viridans streptococcal (VS) experimental (exp) endocarditis (E), abstr. 700, p. 203. In Program and abstracts of the 30th Interscience Conference on Antimicrobial Agents and Chemotherapy. American Society for Microbiology, Washington, D.C. 9. Durack, D. T., and P. B. Beeson. 1972. Experimental bacterial endocarditis. I. Colonisation of a sterile vegetation. Br. J. Exp. Pathol. 53:44–49. 10. Eliopoulos, G. M., and R. C. Moellering. 1991. Laboratory methods used to assess the activity of antimicrobial combinations, p. 434–449. In V. Lorian (ed.), Antibiotics in laboratory medicine. Williams & Wilkins Co., Baltimore. 11. Fantin, B., and C. Carbon. 1990. Importance of the aminoglycoside dosing regimen in the penicillin-netilmicin combination for treatment of Enterococcus faecalis-induced experimental endocarditis. Antimicrob. Agents Chemother. 34:2387–2391. 12. Fluckiger, U., P. Moreillon, J. Blaser, M. Bickle, M. P. Glauser, and P. Francioli. 1994. Simulation of amoxicillin pharmacokinetics in humans for the prevention of streptococcal endocarditis in rats. Antimicrob. Agents Chemother. 38:2846–2849. 13. Flu ¨ckiger, U., C. Segessenmann, and A. U. Gerber. 1991. Integration of pharmacokinetics and pharmacodynamics of imipenem in a human-adapted mouse model. Antimicrob. Agents Chemother. 35:1905–1910. 14. Francioli, P., W. Ruch, D. Stambulian, and the International Infective Endocarditis Study Group. 1993. Treatment of streptococcal endocarditis with a single daily dose of ceftriaxone and netilmicin for 14 days, 11. In Program and abstracts of the 2nd International Symposium on Modern Concepts in Endocarditis. 15. Fuursted, K. 1987. Comparison of the post-antibiotic effect of Streptococcus faecalis and Streptococcus faecium with ampicillin alone or combined with streptomycin: studies on a novel type of antimicrobial interaction. Acta Pathol. Microbiol. Immunol. Scand. Sect. B 95:251–357. 16. Fuursted, K. 1988. Evaluation of the combination effects of ampicillin or vancomycin combined with streptomycin, gentamicin, tobramycin or netilmicin against enterococci. Acta Pathol. Microbiol. Immunol. Scand. 97:23– 26. 17. Garrison, P., and L. R. Freedman. 1970. Experimental endocarditis. I. Staphylococcal endocarditis in rabbits resulting from placement of a polyethylene catheter in right side of the heart. Yale J. Biol. Med. 42:394–410. 17a.Gavalda, J., P. J. Cardona, A. Pahissa, B. Almirante, E. Crespo, L. Pou, J. M. Martinez-Vasquez, J. Steckelberg, M. Rouse, and W. R. Wilson. 1993. Effect of Gentamicin (G) dosing interval on G plus ampicillin (A) treatment (Rx) of enterococcal experimental endocarditis, abstr. 156, p. 148. In Program and abstracts of the 33rd Interscience Conference on Antimicrobial Agents and Chemotherapy. American Society for Microbiology, Washington, D.C. 18. Gavalda `, J., A. Pahissa, B. Almirante, M. Laguarda, E. Crespo, L. Pou, and F. Ferna ´ndez. 1995. Effect of gentamicin dosing interval on therapy of
21.
22. 23.
24.
25.
26.
27.
28. 29.
30. 31. 32. 33.
34. 35.
36.
37.
38.
39.
viridans streptococcal experimental endocarditis with gentamicin plus penicillin. Antimicrob. Agents Chemother. 39:2098–2103. Gerber, A. U., H. P. Brugger, C. Feller, T. Stritzko, and B. Stalder. 1986. Antibiotic therapy of infections due to Pseudomonas aeruginosa in normal and granulocytopenic mice: comparison of murine and human pharmacokinetics. J. Infect. Dis. 153:90–97. Gerber, A. U., S. Kozak, C. Segessenmann, U. Fluckiger, T. Bangerter, and U. Greter. 1989. Once-daily versus thrice-daily administration of netilmicin in combination therapy of Pseudomonas aeruginosa infection in a manadapted neutropenic animal model. Eur. J. Clin. Microbiol. Infect. Dis. 8:233–237. Gerber, A. U., T. Stritzko, C. Segessenmann, and B. Stalder. 1991. Simulation of human pharmacokinetics profiles in mice, and impact on antimicrobial efficacy of netilmicin, ticarcillin and ceftazidime in the peritonitis-septicemia model. Scand. J. Infect. Dis. 74:195–230. Hessen, M. T., P. G. Pistakis, and M. E. Levison. 1989. Postantibiotic effect of penicillin plus gentamicin versus Enterococcus faecalis in vitro and in vivo. Antimicrob. Agents Chemother. 33:608–611. Kapusnik, J. E., C. J. Hackbarth, H. F. Chambers, T. Carpenter, and M. A. Sande. 1988. Single, large, daily dosing versus intermittent dosing of tobramycin for treating experimental Pseudomonas pneumonia. J. Infect. Dis. 158:7–12. Legget, J. E., B. Fantin, S. Ebert, K. Totsuka, W. Calame, H. Mattie, and W. A. Craig. 1989. Comparative antibiotic dose-effect relations at several dosing intervals in murine pneumonitis and thigh-infection models. J. Infect. Dis. 159:281–292. Maller, R., H. Ahrne, T. Eilard, I. Eriksson, I. Lausen, and the Scandinavian Amikacin Once Daily Study Group. 1991. Efficacy and safety of amikacin in systemic infections when given as a single daily dose or in two divided doses. J. Antimicrob. Chemother. 27(Suppl. C):121–128. Mizen, L., G. Woodnutt, I. Kernutt, and E. Catherall. 1989. Simulation of human serum pharmacokinetics of ticarcillin-clavulanic acid and ceftazidime in rabbits, and efficacy against experimental Klebsiella pneumoniae meningitis. Antimicrob. Agents Chemother. 33:693–699. Moellering, R. C., Jr. 1995. Enterococcus species, Streptococcus bovis, and Leuconostoc species, p. 1826–1860. In G. E. Mandell, J. E. Bennet, and R. Dolin (ed.), Principles and practice of infectious diseases, 4th ed. Churchill and Livingstone, Inc., New York. Norrby, S. R., T. O’Reilly, and O. Zak. 1993. Efficacy of antimicrobial agent treatment in relation to treatment regimen: experimental models and clinical evaluation. J. Antimicrob. Chemother. 21(Suppl. D):41–54. Norris, S., C. H. Nightingale, and G. L. Mandell. 1990. Tables of antimicrobial agent pharmacology, p. 434–460. In G. E. Mandell, R. G. Douglas, and J. E. Bennet (ed.), Principles and practice of infectious diseases, 3rd ed. Churchill and Livingstone, Inc., New York. Olier, B., G. Viotte, J. P. Morin, and J. P. Fillastre. 1983. Influence of dosage regimen on experimental tobramycin nephrotoxicity. Chemotherapy (Basel) 29:385–394. Perlman, B. B., and L. R. Freedman. 1971. Experimental endocarditis. II. Staphylococcal infection of the aortic valve following placement of a polyethylene catheter in the left side of the heart. Yale J. Biol. Med. 44:206–213. Prins, J. M., H. R. Bu ¨ller, E. J. Kuijper, R. A. Tange, and P. Speelman. 1993. Once versus thrice daily gentamicin in patients with serious infections. Lancet 341:335–338. Saleh-Mghir, A., A.-C. Cre´mieux, J.-M. Vallois, M. Muffat-Joly, C. Devine, and C. Carbon. 1992. Optimal aminoglycoside dosing regimen for penicillintobramycin synergism in experimental Streptococcus adjacens endocarditis. Antimicrob. Agents Chemother. 36:2403–2407. Sturm, A. W. 1989. Netilmicin in the treatment of gram-negative bacteremia: single daily versus multiple daily dosage. J. Infect. Dis. 159:931–937. ter Braak, E. W., P. J. de Vries, K. P. Bouter, S. G. van der Vegt, G. C. Dorrestein, J. W. Nortier, A. van Dijk, R. P. Verkooyen, and H. A. Verbrugh. 1990. Once-daily dosing regimen for aminoglycoside plus beta-lactam combination therapy of serious bacterial infections: comparative trial with netilmicin plus ceftriaxone. Am. J. Med. 89:58–66. Tulkens, P. M. 1991. Pharmacokinetic and toxicological evaluation of a once-daily regimen versus conventional schedules of netilmicin and amikacin. J. Antimicrob. Chemother. 27(Suppl. C):49–61. Wood, C. A., D. R. Norton, S. J. Kohlhepp, P. W. Kohnen, G. A. Porter, D. C. Houghton, R. E. Brummett, W. M. Bennett, and D. N. Gilbert. 1988. The influence of tobramycin dosage regimen on nephrotoxicity, ototoxicity and antibacterial efficacy in a rat model of subcutaneous abscess. J. Infect. Dis. 158:13–22. Woodnutt, G., V. Berry, and L. Mizen. 1992. Simulation of human serum pharmacokinetics of cefazolin, piperacillin, and BRL 42715 in rats and efficacy against experimental intraperitoneal infections. Antimicrob. Agents Chemother. 36:1427–1431. Woodnutt, G., E. J. Catherall, I. Kernutt, and L. Mizen. 1988. Temocillin efficacy in experimental Klebsiella pneumoniae meningitis after infusion into rabbit plasma to simulate antibiotic concentrations in human serum. Antimicrob. Agents Chemother. 32:1705–1719.
