Combination antibiotic therapy: comparison of constant infusion and intermittent bolus dosing in an experimental animal model

Journal of Antimicrobial Chemotherapy (1985) 15, Suppl. A, 313-321 Combination antibiotic therapy: comparison of constant infusion and intermittent bolus dosing in an experimental animal model Joyce J. Mordenti, Richard Quintiliani and Charles H. Nightingale Department of Pharmacy Services. Division of Infectious Diseases, Hartford Hospital, Hartford, CT 06115, U.S.A. To determine the effect of the mode of administration on antibiotic efficacy, 300 neutropenic rats were infected intrapcritoneally with an LD-70 inoculum of Pseudomonas aeruginosa and treated with synergistic combinations of amikacin and ticartiliin by intermittent or constant infusion technique. The treatment regimens were designed to provide the same peak serum concentrations that would be obscfted in humans receiving these drugs. Drug administration over the 24-h period was controlled to ensure that intermittent and constant infusion techniques achieved the same area under the serum concentration/time curves. Based on cumulative mortality at 96 h and viable bacterial cell counts at the site of inoculation constant infusion of both antibiotics produced the best therapeutic results. Introduction Of the factors responsible for increasing the cancer patient's susceptibility to infection, neutropenia is the most important; the neutropenic patient is unable to mount an adequate inflammatory response and, hence, must rely almost entirely upon antibiotics to control the infection. The administration schedule of drugs in neutropenic patients may be important in determining whether a patient overcomes infection since Gramnegative bacilli recover rapidly from the effects of /Mactam antibiotics once they are eliminated (Toothaker, Welling & Craig, 1982; Bundtzen et al, 1981; Zinner, Husson & Klastersky, 1981; Bodey, Valdivieso & Yap, 1980; Al-Asadi, Greenwood & O'Grady, 1979; Grasso et al., 1978). Comparative trials relating mode of drug administration to therapeutic efficacy in neutropenic patients have shown that continuous antibiotic infusion usually resulted in eradication of infections even when white blood cell counts remained suppressed (Feld et al., 1977; Bodey et al, 1976; Bodey et al, 1975). Clinical and experimental investigations dealing with this question have been limited, and convincing proof or evidence of the superiority of either mode of administration is lacking (Barza, Kane & Baum, 1983; Gerber et al, 1983; Powell et al, 1983; Klastersky, Thys & Mombelli, 1981; Sande et al, 1981; Feld et al, 1979; Issell et al, 1979; Keating et al, 1979; Bergeron, Nguyen & Gauvreau, 1978). To study further the effect of mode of administration on survival from a bacterial infection, neutropenic rats were infected experimentally with Pseudomonas aeruginosa and treated with synergistic 0305-7453/85/15A313+ 09 S02.00/0 313 1985 The British Society for Antimicrobial Chemotherapy

314 J. J. Mordenti et aj. combinations of amikacin and ticarcillin by constant infusion or intermittent dosing schedules. Methods Selection of test organism The strain of Ps. aeruginosa selected for these studies was resistant to amikacin and ticarcillin when used singularly, but sensitive to the combination. This criterion was established to ensure the effects of mode of administration on synergy would be maximized. The minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBQ of amikacin was 40 and 80 mg/l respectively and the MIC and MBC of ticarcillin was 160 and 320 mg/l respectively. The organism was sensitive to synergistic combinations (i MIC of each drug) of these antibiotics. Neutropenic animal model A neutropenic rat model was selected for the present study for reasons of convenience (cost, size, disposition) and general acceptance by clinical researchers for studies of this nature. Female Sprague-Dawley rats (Taconic Farms, German town, NY 12526) weighing 180-220 g were used exclusively. The animals were made neutropenic by administering an 80 mg/kg intraperitoneal dose of cyclophosphamide on day 0 and 60 mg/kg on day 4 of the experiment It was previously determined that this dose produced neutropenia for the entire experimental period (Mordenti, 1983). All animals were injected ip with 1 ml of a 5% gastric mucin solution 30 min prior to injecting 3 x 10 8 micro-organisms in the logarithmic growth phase into the peritoneal cavity. These injections took place one day after the second dose of cyclophosphamide (day 5). This procedure resulted in a 70% animal mortality (LD 70 ). The pharmacokinetics of amikacin and ticarcillin were determined in these animals so that the appropriate dose and dosing frequency could be established. The goal was to provide the same peak serum concentrations that would be obtained in humans, i.e. 30 mg/l for amikacin and 200 mg/l for ticarcillin. Cumulative mortality in treatment trials Three-hundred female Sprague-Dawley rats were made neutropenic with ip cyclophosphamide injections on day 0 and day 4. An LD-70 inoculum of Ps. aeruginosa was injected intraperitoneally on day 5. Antibiotic therapy was instituted 4 h after the initiation of infection and continued for 24 h. Drug was administered by intermittent im injection (every 2 h for amikacin and every 3 h for ticarcillin) or by frequent (every 30 min) im injection to simulate constant infusion therapy. All injections were made in the hind legs, and all possible antibiotic combinations were evaluated. Treatment regimens were as follows: amikacin every 2 h (Ab), amikacin every ^ h (approximating to continuous infusion (Ac)), ticarcillin every 3 h (Tb), ticarcillin every \ h (Tc), both drugs bolus dosing (AbTb), both drugs constant infusion (AcTc) and combinations of one drug infused and one bolus dosed (AcTb, AbTc). Each treatment group was comprised of 30 neutropenic rats; the control group was comprised of 60

Combination antibiotic therapy 315 neutropenic rats. Animal deaths were recorded every 12 h for 96 h. After 96 h, the white blood cell count began returning to normal. At the end of the study, all surviving rats were sacrificed with a 0-5 ml im dose of euthanasia mixture (35 mg/ml ketamine and 7 mg/ml succinylcholine). As a control procedure, it was previously determined that the administration of ip mucin, ip heat killed Ps. aeruginosa, multiple injections of saline (01 ml im q 3 h for 72 h), injections of amikacin (llmg/kgimq 3h for 72 h), ticarcillin (loomg/kgq 3h for 72 h and amikacin + ticarcillin (given as above) for a ten day period of time did not result in any animal deaths. Bacterial quantitation To investigate the effect of mode of administration on bactericidal potency, an additional 50 rats were rendered neutropenic with cyclophosphamide dnd infected lntraperitoneally with an LD-70 inoculum of Ps. aeruginosa. Forty rats were randomly assigned to the following antibiotic treatment regimens: AbTb, AbTc, AcTb, and AcTc. Ten rats served as controls. Following 24 h of antibiotic therapy, all rats were sacrificed with 0-5 ml of euthanasia mixture. Cardiac blood and peritoneal fluid were removed for microbiological assay. The latter was obtained by exposing the peritoneum, injecting 20 ml of sterile normal saline and 10 ml air, shaking the animal vigorously and withdrawing approximately 20 ml of fluid. Two millilitres of blood were obtained by cardiac puncture. The blood and peritoneal samples were serially diluted in sterile normal saline for enumeration by microbiological plate counts. Statistical evaluation The Chi-square test was used to establish that a statistically significant difference existed in comparisons of differences in proportions. The Chi-square test with Yates' correction was used for pair-wise comparison of differences in proportions. The Fisher's exact test was used for pair-wise comparison of differences in proportions whenever the excepted frequencies for the Chi-square test were less than 5. A P value less than or equal to 0-05 was considered significant (Riegelman, 1981; Colton, 1974). The F test for analysis of variance was used to establish that a statistically significant difference existed in comparisons of differences in geometric means. A P value less than or equal to 0-05 was considered significant. When a statistically significant difference was found, Fischer's least significant difference test for pair-wise comparisons was used to determine which comparisons made contributions to the overall difference. A P value less than or equal to 0-01 was considered significant (Riegelman, 1981; Colton, 1974; Snedecor & Cochran, 1974). Antibiotic pharmacokinetic studies Results When rats were handled in a manner which minimized stress, the in-vivo antibiotic serum profiles for ticarcillin and amikacin could be described by a one compartment open model. Ticarcillin (loomg/kg) produced a peak concentration of 180-200 mg/1; the serum half-life was approximately 15min. Amikacin (11 mg/kg) produced a peak concentration of 28-30 mg/1; the serum half-life was approximately 17 mm.

316 J. J. Mordenti et al. IOOO 100-100 200 300 Time (mm) 400 500 Figure 1. Ticarcillin pharmacokinetics in two representative rats after intramuscular injections every 3 h. The curves are theoretical and the points experimental The rodent pharmacokinetic parameters were used to design proper multiple dose administration schedules, i.e., intramuscular injections of amikacin every 2 h (q 2 h) and ticarcillin every 3 h (q 3 h). Constant infusion was approximated by intramuscular injections every half hour (q ^ h). The log of the antibiotic serum concentration was plotted as a function of time for each multiple dosing experiment (Figures 1 to 4). In each trial, the experimental serum concentrations compared favourably with the expected pharmacokinetic curves. Cumulative mortality The results (Figure 5) show that those treatment regimens which had drug infused resulted in higher survival rates the best results were obtained when both drugs were infused. Survival among rats treated with amikacin alone (Ab or Ac) or ticarcillin bolus dosing (Tb) was not significantly different than the survival of control rats. Survival among rats treated with ticarcillin continuous infusion (Tc) or combination drug therapy (AbTb, AbTc, AcTb, and AcTc) was significantly better than controls (P < 0-05 for Tc, and P < 0-01 for combination therapy). Combination drug therapy consisting of continuous infusions of both drugs (AcTc) was significantly better (i ) <O05) than combination therapy in which the ticarcillin was administered intermittently (AbTb or AcTb).

Combination antibiotic therapy 317 1000 100-100 200 300 Time (mm) Fignre 2. Ticarcillin pharmacokinetics in two representative rats after intramuscular injections every $ h. The curves are theoretical and the points experimental Bacterial quantitation The results of quantitative cultures of peritoneal lavage obtained 24 h after challenge with P. aeruginosa are shown in Table I. The bacterial counts correlated with cumulative mortality, and the combination of constant infusion amikacin plus constant infusion ticarcillin was associated with the smallest number of bacteria in the peritoneal cavity. A one-way analysis of variance confirmed a statistical difference in the log colonyforming units (cfu) found in rats in the various treatment groups (P < 0-05). The low number of bacteria found in rats in the AcTc and AcTb treatment groups was responsible for this significant difference. Discussion Presently the most popular therapy for suspected sepsis in neutropenic patients is a combination of an anti-pseudomonas penicillin derivative (e.g., carbenicillin, ticarcillin, mezlocillin, piperacillin) and an anti-pseudomonas aminoglycoside (e.g., gentamicin, tobramycin, amikacin) (Bodey, 1983; KJastersky, 1983). There has been reluctance to employ monotherapy with a /Mactam antibiotic or an aminoglycoside owing to the clinical observation that synergistic combinations against the usual Gram-negative bacteraemic organisms (e.g., Klebsiella, E. coli, Ps. aeruginosa) have been associated with higher survival rates than those obtained with single drug therapy, even against susceptible isolates (Klastersky, Meunier-Carpentier & Prevost, 1977) 400

318 J. J. Mordeoti el al. 100 100 200 300 Tim* (min) 400 300 Figure 3. Amikacin pharmacokinetics in two representative rats after intramuscular injections every 2 h The curves are theoretical and the points experimental The mode of antibiotic administration (constant infusion vs. intermittent injection) is currently a subject of debate; for instance, there are several clinical studies in which the continuous infusion of antibiotics appeared to be somewhat more efficacious than that of intermittent dosing in neutropenic patients, yet constant infusion therapy is rarely employed in these patients (Feld, 1977; Valdivieso et al, 1974). The ability of certain organisms to recover quickly from the effects of antibiotics was demonstrated in in-vitro studies when Gram-negative bacteria sensitive to carbenicillin and cephalothin regrew promptly once the antibiotic was depleted (Robinson, 1973). This absence of socalled post-antibiotic effect in the treatment of Gram-negative bacteria with /Mactam antibiotics may have particular clinical relevance in the neutropenic patient owing to their suboptimal natural host defences against bacterial proliferation. As suggested by Keating et al. (1979), for the best clinical outcome in patients with serious defects in host defences, constant inhibitory concentrations of antibiotics may be needed in serum or tissue to protect against the rapid growth of bacteria that might occur if antibiotic levels are absent at these sites for prolonged periods. Although we found in this investigation that the dual bolus dosing schedule produced high peak antibiotic serum concentration for amikacin and ticarcillin, the peaks did not coincide at every dosing interval, producing intervals when only one drug was detectable in the blood and trough concentrations which were low for both drugs. Overall, the time of exposure to the synergistic effect of both antibiotics was

Combination antibiotic therapy 319 100 100 200 Tims (min) Figure 4. Amikacin phaimacolunetics in two representative rats after intramuscular injections every \ h. The curves are theoretical and the points experimental 300 Figure 5. Cumulative mortality (per cent) for antibiotic treatment tnals: Ac, amikacin 'continuous infusion' (CI); Ab, amikaan bolus dosing every 2 h (bd); Tc, ticarcilbn CI; Tb, ticarcillin bd every 3 h, AbTb, both drugs bd, AcTc, both drags CI, AbTc, AcTb, combination of bd and CI.

320 J. J. Mordenti et al. Table L Efficacy of various dosage regimens of amikacin and ticarcillin based on bacterial recovery in neutropenic rats sacrificed at 24 h Dosage regimen Control AbTb AcTb AbTc AcTc No. of rats positive/ no. tested Peritoneal fluid 10/10' 8/9' 8/9' 8/9' 6/9' Log cfu±sd. ll-4±0-3 59±O76 51 ±0-77 5-8 ±0-52 4-8±0-ll ' AU control rats died before 24 h. The log cfu represents the bacterial count at the time of death * One rat died before 24 h, and the bacterial count for the rat is not included in the tabulation. The log cfu in treatment rats that died during therapy were as follows AbTb, 8-6; AcTb, 1O4; AbTc, 7-8; AcTc, 6-1. minimal and is probably the cause of the poor response associated with this mode of therapy. Combination therapy consisting of bolus dosing of one drug plus continuous infusion of the second drug was of intermediate success. There were no drug free intervals: one antibiotic was always at J MIC, and the other antibiotic had substantial peaks above \ MIC and troughs well below i MIC. The exposure to the synergistic combination was greater than with the dual bolus dosing schedule but less than with the dual continuous infusion schedule. Intramuscular injections of both antibiotics every { h produced steady-state serum concentrations of ^ MIC for each drug throughout the dosing interval. This concentration represents the 'synergy' MIC for the infecting organism. Although this dosage regimen lacked high peak antibiotic concentrations for each drug, it afforded maximal protection by not allowing prolonged antibiotic-free periods to occur and by achieving the maximal time for a synergistic interaction to occur. In summary, our data indicate that constant infusion of both an aminoglycoside and a penicillin-derivative achieved the best results. When levels were sustained well above the MIC and MBC of the antibiotics for Ps. aeruginosa, which was the situation with constant infusion of antibiotics, animals survived longer and bacteria were eliminated more rapidly. It must be stressed that this is especially true of antibiotic combinations that depend upon the development of synergy in order to eradicate the organism. Intermittent bolus dosing tends to result in non-synergistic drug concentrations for a portion of the dosing interval. The greater the non-synergy the less effective the therapy. Obviously if the organism is exquisitely sensitive to the drug or drug combinations, i.e. has extremely low MIC values, the mode of administration will be relatively unimportant since adequate coverage will result with either mode. Likewise, differences in mode of administration may not be critical in patients with normal host defence mechanisms since organism 'killing' can occur for a period of time after antibiotic concentration fall below the MIC. In neutropenic patients, especially in infections with organisms that have only moderate or poor susceptibility to the drug(s),

Combination antibiotic therapy 321 the mode of administration may be important, and the clinician should try to increase the periods of inhibitory drug concentrations by using constant antibiotic administration. Acknowledgements This project was supported by a generous grant from the Combined Hospital Fund, Hartford, CT 06115. We wish to thank Al Beroth, Bill Dyckman, Ed Hall, Nancy Reichart, and John Fakas from the Hartford Hospital Animal Research Facility for providing technical assistance and excellent animal care throughout these studies. This work was abstracted from Joyce Mordenti, "Combination antibiotic therapy: Comparison of constant infusion and intermittent bolus dosing in an in-vitro kinetic model and an experimental animal model", Ph.D. dissertation, University of Connecticut, 1983. Copies are available through University Microfilm International, 300 N. Zeeb Road, Ann Arbor, MI 48106, U.S.A. References Al-Asadi, M., Greenwood, D. & O'Grady, F. (1979) In vitro model simulating the form of exposure of bacteria to antimicrobial drugs encountered in infection. Antimicrobial Agents and Chemotherapy 16, 77-80. Barza, M., Kane, A. & Baum, J. (1983). Comparison of the effects of continuous and intermittent systemic administration on the penetration of gentamicin into infected rabbit eyes. Journal of Infectious Diseases 147, 144-8. Bergeron, M., Nguyen, B & Gauvreau, L. (1978). Influence of constant infusion versus bolus injection of antibiotics on in vivo synergy. Infection, 6, Suppl. I, S38-S46. Bodey, G (1977). Infectious complications in the cancer patient Current Problems in Cancer 1, 3-63. Bodey, G., Bolivar, V, Fainstein, L. & Jadeja, L. (1983). Infections caused by Pseudomonas aerugutosa. Reviews of Infectious Diseases 5, 279-313. Bodey, G., Chang, H., Rodriguez, V. & Stewart, D. (1975). Feasibility of administering aminoglycoside antibiotics by continuous intravenous infusion. Antimicrobial Agents and Chemotherapy 8, 328-33. Bodey, G., Rodriguez, V., Valdivieso, M. & Feld, R. (1976). Amikacin for treatment of infections in patients with malignant diseases. Journal of Infectious Diseases 134, Suppl., S241-S427. Bodey, G., Valdivieso, M. & Yap, B. (1980) The role of schedule in antibiotic therapy of the neutropenic patient Infection 8, Suppl. 1, S75-S81. Bundtzen, R., Gerber, A., Cohn, D. & Craig, W (1981) Postantibiotic suppression of bacterial growth. Review of Infectious Diseases 3, 28-37. Colton, T. (1974). Statistics in Medicine, pp. 151-88. Little, Brown & Co, Boston. Feld, R., Valdivieso, M., Bodey, G. & Rodriguez, V. (1977). A comparative trial of sisomicin therapy by intermittent versus continuous infusion American Journal of Medical Science 11 A, 179-88. Feld, R., Tuffnell, P. el al. (1979). Empiric therapy for infections in granulocytopenic cancer patients: Continuous infusion of amikacin plus cephalothin. Archives of Internal Medicine 139, 310-^t. Gerber, A., Craig, W. el al. (1983) Impact of dosing intervals on activity of gentamian and ticarcillin against Pseudomonas aeruginosa in granulocytopenic mice Journal of Infectious Diseases 147, 910-7. Grasso, S., Meinardi, G decarneri, I. & Tamassia, V. (1978). New in vitro model to study the effect of antibiotic concentration and rate of elimination on antibacterial activity Antimicrobial Agents and Chemotherapy 13, 570-6

322 J. J. Mordenti et al. Issell, B., Keating, M., Valdivieso, M. & Bodey, G. (1979) Continuous infusion tobramycin combined with carbenicilhn for infections in cancer patients. American Journal of Medical Sciences 277,311-8. Keating, M., Bodey, G., Valdivieso, M. & Rodriguez, V. (1979). A randomized comparative trial of three aminoglycosides comparisons of continuous infusions of gentamicin, amikacin, and sisomicin combined with carbenicillin in the treatment of infections in neutropenic patients with malignancies. Medicine 58, 159-70. Klastersky, J. (1983). Empiric treatment of infections in neutropenic patients with cancer. Review of Infectious Diseases. Suppl. I. 5, 521-31. Klastersky, J., Meunier-Carpentier, F. & Prevost, J. (1977). Significance of antimicrobial synergism for the outcome of gram-negative sepsis. American Journal of Medical Science 273, 157-67. Klastersky, J., Thys, J. & Mombelli, G. (1981). Comparative studies of intermittent and continuous administration of aminoglycosides in the treatment of bronchopulmonary infections due to gram-negative bacteria. Review of Infectious Diseases 3, 74-83. Mordenti, J. (1983). Combination antibiotic therapy: comparison of constant infusion and intermittent bolus dosing in an in vitro kinetic model and an experimental animal model. Ph.D. dissertation. University of Connecticut. Powell, S., Thompson, W., Lythe, M. et al. (1983). Once daily vs. continuous aminoglycoside dosing: Efficacy and toxicity in animal and clinical studies of gentamicin, netilmicin, and tobramycin. Journal of Infectious Diseases 147, 918-32. Riegelman, R. (1981). Studying a Study and Testing a Test, pp. 209-22. Little, Brown & Co., Boston. Robinson, G. (1973). Plasma concentrations of penicillin in relation to the antibacterial effect. In Biological Effects of Drugs in Relation to their Concentrations (Davies, D S. & Priehard, B., Eds), pp. 183-9. MacMillan Press, N.Y. Sande, M., Korzeniowski, O. et al. (1981). Intermittent or continuous therapy of experimental meningitis due to Streptococcus pneumoniae in rabbits preliminary observations on the post-antibiotic effect in vivo. Review of Infectious Diseases 3, 98-109 Snedecor, G. & Cochran, W. (1974). Statistical Methods, 6th Ed., pp 258-98. Iowa State University Press, Ames, Iowa. Toothaker, R., Welling, P. & Craig, W. (1982). An in vitro model for the study of antibacterial dosage regimen design. Journal of Pharmaceutical Science 71, 861-4. Valdivieso, M., Feld, R., Rodriguez, V. Bodey, G. (1974). Amikacin therapy of infections in neutropenic patients. American Journal of Medical Science 268, 149 Zinner, S., Husson, M & Klastersky, J (1981). An artificial capillary in vitro kinetic model of antibiotic bactericidal activity. Journal of Infectious Diseases 144, 583-7