Cerebrospinal Fluid Penetration of Imipenem and Cilastatin (Primaxin) in Children with Central Nervous System Infections

Transcription

1 ANTIMIWROBIAL AGENTS AND CHEMOTHERAPY, Apr. 1986, p /86/ $02.00/0 Copyright X 1986, American Society for Microbiology Vol. 29, No. 4 Cerebrospinal Fluid Penetration of Imipenem and Cilastatin (Primaxin) in Children with Central Nervous System Infections RICHARD F. JACOBS,'* GREGORY L. KEARNS, 2 ALLEN L. BROWN,"3 AND DARRYL C. LONGEE' Departments of Pediatrics,1 Pharmaceutics,2 and Pharmacy Practice,3 University of Arkansas for Medical Sciences and Arkansas Children's Hospital, Little Rock, Arkansas Received 28 October 1985/Accepted 22 January 1986 Cerebrospinal flbid (CSF) penetration of imipenem-cilastatin was evaluated in 20 children (aged 4 months to 11 years) with central nervous system infections. A total of 10 children received a single 25-mg/kg intravenous dose, and 10 received three 25-mg/kg intravenous doses at 6-h intervals. Blood and CSF were obtained 1.5 to 2.5 h after the last dose during the early (days 1 to 3) and the late (days 7 to 10) stages of infection. Imipenem concentrations after single-dose infusion in serum and CSF during the early phase of treatment (8.59 ± 0.95 and ,ig/ml, respectively) were similar to those during the late phase (9.96 ± 2.36 and 2.08 ± 1.14,ug/ml, respectively). Concentrations of imipenem in serum and CSF after multiple-dose infusion during the early phase (11.97 ± 2.03 and 1.87 ± 0.29,ug/ml, respectively) were similar to those during the late phase ( and 1.22 ± 0.11 p.g/ml, respectively). There were no significant differences in the serum or CSF imipenem concentrations between the single- and multiple-dose groups during the early or late treatment stages. Cilastatin concentratiods in serum and CSF were similar in all groups with the exception of the multiple-dose, early- versus late-phase evaluationt of CSF cilastatin concentration. There was no correlation between age or absolute. CSF neutrophil count and the serum or CSF concentrations of imipenem or cilastatin. We found a mean CSF penetration of 15 to 27% for imipenem and 16 to 66% for cilastatin in children. These findings suggest that imipenem-cilastatin sufficiently penetrates into CSF in children to warrant further investigation of this compound in pediatric central nervous system infections. Primaxin (Merck Sharp & Do4me Laboratories, West Point, Pa.) is a 1:1 combination of a stable derivative of thienamycin, N-formimidoyl thienamycin (imipenem), and the renal dihydropeptidase inhibitor cilastatin (10). Imipenem has a broad antibacterial spectrum that includes aerobic and anaerobic gram-positive and gram-negative microorganisms. The compound has documented activity against Staphylococcus aureus, methicillin-resistant S. aureus, Staphylococcus epidermidis, Pseudomonas sp., and Listeria monocytogenes (8, 11, 13, 18). Previous work has shown that intravenous administration of 25 mg of imipenem-cilastatin (1:1) per kg at 6-h dosing intervals provides therapeutic serum levels of imipenem in children without evidence of accumulation (9). The purpose of this investigation was to determine the penetration of imipenem and cilastatin into cerebrospinal fluid (CSF) of children with central nervous system infections. Assessment was performed after single- and multiple-dose administration during the first and last 3 days of parenteral antibiotic therapy to determine the effect of resolution of infection on the penetration of both compounds into CSF. (Part of this research was presented at the 25th Interscience Conference on Antimicrobial Agents and Chemotherapy, Minneapolis, Minn., 29 September to 2 October 1985.) MATERIALS AND METHODS Patient selection. Twenty children (ages 4 months to 11 years) who were admitted to the Arkansas Children's Hospital with culture-proven bacterial central nervous system infections made up the study population. A total of 3 patients had infections associated with ventriculo-peritoneal shunts, and 17 had bacterial meningitis. The genders of subjects in the single-dose (four male, five female) and * Corresponding author. multiple-dose (five male, four female) groups were compar. rable. The mean age of subjects in the single-dose group (52.0 ± months) and, accordingly, their total body weight (20.1 ± 4.55 kg) were significantly greater than those for subjects in the multiple-dose group (9.11 ± 1.46 months and 8.15 ± 0.85 kg, respectively). The absolute neutrophil counts in the initial CSF samples from both the single- and multiple-dose groups were comparable (2,857 ± 2,282.7 versus 2,719.3 ± 1,089.7 cells per mm3, respectively), although considerable intersubject variability was evident for both groups. Patients were prospectively randomized to receive either sihgle or multiple (three) 25-mg/kg doses of imipenem-cilastatin intravenously. The patients received the single- or multiple-dose regimen during the first 3 days (early stage) and again during the last 3 days (late stage) of concomitant intravenous antibiotic therapy consisting of cefotaxime (200 mg/kg per day) with or without ampicillin (200 to 400 mg/kg per day). Patients were excluded from the study if they were critically ill, hemodynamically unstable, had abnormal renal or hepatic function, were immunocompromised, or had cystic fibrosis. The study was reviewed and approved by the Institutional Review Board for Human Experimentation. Informed parental consent and, when possible, patient assent, were obtained before subjects were enrolled. Microbiology and antibiotic susceptibility. The microbiologic demographics for patients in the single-dose (three ventriculo-peritoneal shunt infections [one S. aureus and two S. epidermidis] and seven patients with bacterial meningitis [three with Haemophilus influenzae type b, two with Streptococcus pneumoniae, and two with Neisseria mettingitidis]) and multiple-dose groups (10 patients with bacterial meningitis [six H. influenzae, type b, two N. meningitidis, one S. pneumoniae, and one Escherichia coli]) were similar, with H. influenzae being the predominant pathogen. The 670

2 VOL. 29, 1986 IMIPENEM-CILASTATIN IN CSF 671 TABLE 1. Microbiology and antibiotic susceptibility results No. of Mean MIC (pg/ml) + Mean MBC (,ug/ml) + Mean CSF bactericidal titer ± SD' (range) for: Organism cases SD' (range) SD' (range) Day 1-3 Day 7-10 H. influenzae, type b ± 0.15 ( ) 0.26 ± 0.34 ( ) 1: (1:16-1:512) 1:73 ± 77 (1:8-1:256) S. pneumoniae ± 0.12 ( ) ( ) 1: (1:64-1:128) 1:35 ± 28 (1:8-1:64) N. meningitidis ± 0.03 ( ) ( ) 1: (1:8-1:32) 1: (1:8-1:16) E. coli <1:8 1:16 S. aureus :8 NDb S. epidermidis ± 0.7 (1-2) 3.0 ± 1.4 (2-4) 1:8 1:8 a Where applicable. b ND, Not done. antibiotic susceptibility testing (MIC and MBC) for all CSF isolates was performed against imipenem by using a standard microtiter susceptibility technique (7). Susceptibility of the CSF pathogens to serial dilutions of imipenem was performed by using a standard 105 inoculum, Mueller-Hinton broth plus Supplement C for H. influenzae and N. meningitidis, and Todd-Hewitt broth for S. pneumoniae. Tryptic soy broth was used for the S. aureus and S. epidermidis strains. Subcultures for MBC determination were inoculated on 5% sheep blood agar, chocolate agar, or tryptic soy agar for the above isolates. CSF bactericidal titers (Table 1) were calculated for all patients during days 1 through 3 (early stage) and 7 through 10 (late stage) of therapy by using serial dilutions of the patient's CSF (1:2 to 1:1,024) against a standard 105 inoculum of the autologous CSF isolate. The CSF bactericidal titer assays were performed on the same media that were used in the MIC and MBC assays. The values were determined by using previously described methods (15). Drug administration. Imipenem-cilastatin (500 mg) was reconstituted in 0.9% sodium chloride as previously described (9). Study doses were administered intravenously over a 15- to 20-min period within 8 h of preparation by using a constant infusion pump (IVAC 630 pump, IVAC Corp., San Diego, Calif.). During and within 1 h of this time, no other drug was administered ipsilaterally. Specimen collection and preparation. At approximately 1.5 to 2.5 h after complete infusion of the respective study doses, paired blood and CSF samples were obtained via venous sampling from a contralateral extremity and lumbar puncture, respectively. Paired blood and CSF samples were at all times obtained within 10 to 15 min of each other, and exact postinfusion times were recorded. The blood (2.5 ml) and CSF (2.0 ml) samples were collected in glass tubes which did not contain anticoagulant and were transported immediately on ice to the laboratory. Serum was separated from blood samples after centrifugation at 4 C and 2,000 x g for 10 min within 1.0 h of collection. In a similar manner, CSF samples were centrifuged, and the clear supernatant (1.0 to 1.5 ml) was removed. Sample processing. An aliquot of both serum and CSF (0.5 ml) from each sample was added to 0.5 ml of stabilizing solution containing 1 volume of morpholinoethane-sulfonate buffer (MES buffer; Calbiochem-Behring, La Jolla, Calif.) and 1 volume of ethylene glycol (J. T. Baker Chemical Co., Phillipsburg, N.J.). The mixtures were thoroughly vortexed, and the resultant solutions were immediately transferred to two shipping vials and were quick-frozen at -70 C. Frozen samples were shipped on dry ice to Merck Sharp and Dohme Laboratories (West Point, Pa.) for quantitation of imipenem and cilastatin from plasma and CSF. Quantitation of imipenem and cilastatin. Quantitation of imipenem and cilastatin from serum and CSF was accomplished by using high-performance liquid chromatography as previously described (12). The analytical method had a range of linearity of 0.5 to 50.0,ug/ml for both imipenem and cilastatin in serum and CSF, with a lower limit of detection of 0.5,ug/ml for both compounds. The between-day and within-day coefficient of variation for replication was consistently <7% for low (0.5,ug/ml) and high (50.0 jxg/ml) standards from serum and CSF. Analytical recoveries of both imipenem and cilastatin were consistently >90% from each biologic matrix. Frozen stability of imipenem and cilastatin in serum and CSF at -70 C was documented for 90 days. All standard curves for imipenem and cilastatin quantitation were prepared in drug-free samples of serum and serum ultrafiltrate (for quantitation from CSF). Analysis of specimens was performed in triplicate with the mean value reported as the final concentration of either imipenem or cilastatin in serum or CSF. Safety and tolerance. No adverse effects were detected in the patients during the early- or late-stage evaluation periods. The following laboratory studies were monitored both before and 5 days after imipenem-cilastatin administration: hematologic profile, platelet count, blood urea nitrogen, serum creatinine, serum glutamic oxalacetic transaminase, total bilirubin, and alkaline phosphatase. No abnormalities in the aforementioned studies were documented, and no clinical evidence of phlebitis was observed in any subject. Statistical analysis. All results pertaining to grouped demographic, microbiologic, or pharmacologic data were expressed as a mean and a range about the mean. Covariance determinations between demographic data and pharmacologic and microbiologic parameters were performed by using linear least-squares regression analysis. The following tests were conducted in the regression analysis: (i) Student's t test, the null hypothesis (H.) being that the slope of the regression line was zero; (ii) correlation coefficient (r), Ho being that x and y were independent variables; and (iii) analysis of variance, Ho being that the regression of y on x was not linear. Analyses to detect statistically significant differences between the experimental populations for a given variable were conducted by using the appropriate version of

3 672 JACOBS ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 2. CSF and serum concentration data for imipenem Ts' (min) Concnsb (mg/liter) TCSFa (min) ConcncsFb (mg/liter) CSF-to-serum ratio Dose Early Late Early Late Early Late Early Late Early Late Single X SE Range nc Multiple X d SE Range n a TS, Time from end of drug administration to serum sampling time; TCSF, time from end of drug administration to CSF sampling time. b Concns, Serum concentration; ConcnCSF, concentration in CSF. c n = Number of samples. d p < 0.05 for early versus late evaluation. No other significant differences were found between the single- and multiple-dose groups or between the respective early and late evaluations for a given parameter. the two-tailed Student's t test. In all statistical analyses, the acceptable level of significance was an a of All statistical analyses were performed by previously described standard methods (1, 6, 17). RESULTS The microbiologic data for subjects in the single- and multiple-dose groups are summarized in Table 1. Examination of the MBC for each pathogen in the single- and multiple-dose groups (Table 1) revealed values <1.0,ug/ml for all isolates with the exception of one strain each of S. epidermidis (MBC = 2.0,ug/ml), S. aureus (MBC = 4.0,ug/ml), and H. influenzae (MBC = 1.0,ug/ml). The most sensitive organisms from both treatment groups were H. influenzae (MBC = mg/liter) and N. meningitidis (MBC = 0.12 ± 0.02 mg/liter). Excluding the shunt infections caused by S. aureus and S. epidermidis and a meningitis caused by E. coli, the CSF concentrations of imipenem at the early evaluation point exceeded the respective MBC values by 2.8- to 320-fold. The CSF and serum concentration data for imipenem are summarized in Table 2. Examination of the time between drug administration and the sampling of blood and CSF revealed a significantly lower interval between the early (124.6 ± 6.5 min) and late (149.7 ± 8.8 min) treatment phase blood sampling times for the multiple-dose group only. No other significant differences were found between the singleand multiple-dose groups or between the respective early and late treatment phase evaluations. The concentrations of imipenem in serum and CSF were not significantly different between the early and late treatment phases within the single- and multiple-dose groups. No significant differences were found for these parameters between the single- and multiple-dose groups when compared at both the early and late evaluation periods (Table 2). Similarly, when the CSF-to-serum concentration ratios for imipenem were examined, no significant differences were found between the early and late evaluation intervals or between the single- and multiple-dose groups (Table 2). The CSF and serum concentration data for cilastatin are summarized in Table 3. For the single-dose group, there was no difference in serum or CSF concentrations between the early- and late-stage treatment evaluations. The mean CSFto-serum concentration ratio between the early ( ; range, 0.06 to 0.35) and late ( ; range, 0.3 to 1.41) treatment stage evaluations in the single-dose group was significantly (P < 0.05) different (although this difference was skewed by a ratio of 1.41 for a subject at the late evaluation stage). For the multiple-dose group, the CSF cilastatin concentration for the late evaluation stage (1.04 ± TABLE 3. CSF and serum concentration data for cilastatin Ts' (min) Concnsb (mg/liter) TCSFa (min) ConcncsFb (mg/liter) CSF-to-serum ratio Dose Early Late Early Late Early Late Early Late Early Late Single X c SE Range nd Multiple X c e SE Range n a TS, Time from end of drug administration to serum sampling time; TCSF, time from end of drug administration to CSF sampling time. b Concns, Serum concentration; ConcncsF, concentration in CSF. c Significant (P < 0.05) difference between early and late evaluations for respective parameters. d n = Number of samples. I Significant (P < 0.05) difference between single and multiple dose values.

4 VOL. 29, ,ug/ml) was significantly less than the value observed at the early evaluation point (1.59 ± 0.21,ug/ml) despite the fact that the early-versus-late-stage CSF-to-serum ratios were not different. With the exception of the CSF-to-serum ratio at the late-stage evaluation, there were no differences between the single- and multiple-dose groups for the serum or CSF concentrations of cilastatin. With the exception of the late-stage CSF cilastatin concentration after multiple-dose administration (Table 3), there was no apparent accumulation or loss of cilastatin in serum or CSF consequent to changes in meningitis or multiple-dose administration of imipenem-cilastatin. A series of covariance analyses were conducted between both subject age and absolute CSF neutrophil count, and each of the following parameters was calculated at the early-stage evaluation interval for both the single- and multiple-dose groups: (i) serum concentration of imipenem and cilastatin, (ii) CSF concentration of imipenem and cilastatin, and (iii) CSF-to-serum concentration ratio of imipenem and cilastatin. No significant linear relationships were observed between age or absolute CSF neutrophil count and any of these parameters. Additionally, correlations between the serum and CSF concentrations of either imipenem or cilastatin could not be demonstrated when examined for the single- and multiple-dose groups or for the early and late treatment phase evaluations. DISCUSSION In our study, we examined the penetration of imipenem and cilastatin from blood into CSF in a population of infants and children with documented bacterial meningitis or ventriculitis or both at 1.5 to 2.5 h after an intravenous dose of imipenem-cilastatin. We could not demonstrate a correlation between age or absolute CSF neutrophil count on admission and either imipenem or cilastatin concentrations in serum or CSF, or the degree of penetration when examined as the CSF-to-serum concentration ratios for the respective compounds. Imipenem and cilastatin have been characterized as possessing a high degree (18 to 31% of simultaneous plasma concentration) of penetration into CSF in humans (internal communication; Merck Sharp & Dohme Laboratories, West Point, Pa.; 1985) and in animals (14). Chow et al. (5) have reported a 6.0% mean penetration of imipenem into the CSF of normal rabbits calculated by comparison of serum-to-csf area under the concentration-time curves after administration of a 1:1 combination of imipenem-cilastatin. In a study of experimental L. monocytogenes and E. coli meningitis in rabbits, Shibl et al. demonstrated an approximate mean CSF penetration of 22% for imipenem from blood (A. M. Shibl, C. Hackbarth, K. Scott, T. Carpenter, and M. A. Sande, Program Abstr. 25th Intersci. Conf. Antimicrob. Agents Chemother., abstr. no. 688, 1985). We found a mean penetration for imipenem of between 15 and 27% in our patients at 1.5 to 2.5 h after administration of either a single or multiple intravenous dose of imipenem-cilastatin. Furthermore, the extent of CSF penetration of imipenem did not differ between evaluations in the early and late stages of treatment or between the single- and multiple-dose groups (Table 2). Our evaluation of CSF penetration of imipenem and cilastatin at 1.5 to 2.5 h after an intravenous dose of imipenem-cilastatin should have represented post-peak serum concentrations (9). In view of pharmacokinetic similarities between imipenem (9) and ceftazidime (2) in children, we would expect the CSF concentrations of imipenem at 1.5 IMIPENEM-CILASTATIN IN CSF 673 to 2.5 h postdose to approach peak CSF concentration values. These are important considerations, since accumulation of drugs in the CSF can occur consequent to slower elimination from CSF when compared with serum (3, 16). Despite the logistic and ethical limitations entailed in the selection of a single CSF sampling point, we were unable to demonstrate significant accumulation of either imipenem (Table 2) or cilastatin (Table 3) in serum or CSF when the single- and multiple-dose groups were compared. This finding is not in agreement with a recent case report (4) which speculates that accumulation of imipenem in CSF was a possible etiology of seizures in two elderly adults with abnormal renal function who received multiple intravenous doses of imipenem-cilastatin. To our knowledge, the penetration of cilastatin from blood into CSF has not been previously reported in children. We found mean values for CSF penetration of 16 to 66% after single-dose administration of imipenem-cilastatin in the early and late treatment phases, respectively. After multiple-dose administration in the early and late treatment phases, we found CSF-to-serum ratios for cilastatin of 29 and 21%, respectively (Table 3). For both the single- and multiple-dose evaluations, wide ranges in the extent of cilastatin CSF penetration were observed. Despite this variability, we could not demonstrate a significant accumulation of cilastatin in CSF or serum associated with single- or multiple-dose imipenem-cilastatin administration, or between the early and late evaluation phases in the respective treatment groups. Our data suggest that imipenem and cilastatin appear to penetrate from blood into CSF to similar extents after single and multiple doses of intravenous imipenem-cilastatin when examined at 1.5 to 2.5 h after drug administration. Satisfactory bactericidal titers (1:8 to 1:256) were demonstrated in the CSF of all patients for imipenem at both the early and late treatment phases after single- and multiple-dose imipenem-cilastatin administration. With the exception of an apparent increase for the CSF-to-serum ratio of cilastatin in the late treatment phase after single-dose administration (Table 3), the extent of CSF penetration of either imipenem or cilastatin did not appear to be markedly altered during the treatment course for bacterial meningitis or ventriculitis in children. This finding and the lack of correlation between the absolute neutrophil count in the CSF and the CSF concentrations of imipenem or cilastatin suggest that CSF penetration of both agents may not be markedly influenced by the degree of meningeal inflammation. ACKNOWLEDGMENTS R. F. Jacobs is an EL Trudeau Scholar of the American Lung Association. This work was supported in part by a grant from Merck Sharp & Dohme Laboratories. We thank Vincent Ahonkhai for his assistance with analytical support, Ann Augustine for her excellent technical assistance, and Charlotte Bornemeier for her assistance in the preparation of our manuscript. LITERATURE CITED 1. Batson, H. D An introduction to statistics in the medical sciences, p Burgess Publishing Co., Minneapolis. 2. Blumer, J. L., S. C. Aronoff, C. M. Myers, C. A. O'Brien, J. D. Klinger, and M. D. Reed Pharmacokinetics and cerebrospinal fluid penetration of ceftazidime in children with meningitis. Dev. Pharmacol. Ther. 8: Bonati, M., J. Kanto, and G. Tognoni Clinical pharmacokinetics of cerebrospinal fluid. Clin. Pharmacokinet.

5 674 JACOBS ET AL. 7: Brotherton, T. J., and R. L. Kelber Seizure-like activity associated with imipenem. Clin. Pharm. 3: Chow, A. W., K. R. Finlay, H. G. Stiver, and C. L. Carlson Intrathecal penetration of N-formimidoyl thienamycin in normal rabbits: potentiation by coadministration of renal dipeptidase enzyme inhibitor. Antimicrob. Agents Chemother. 23: Dixon, W. J., and F. J. Massey Introduction to statistical analysis, p McGraw-Hill Book Co., New York. 7. Gavan, T. L., and A. L. Barry Microdilution test procedures, p In E. H. Lennette, A. Balows, W. J. Hausler, Jr., and J. P. Truant (ed.), Manual for clinical microbioloy. American Society for Microbiology, Washington, D. C. 8. Gutierrez-Nufiez, J., P. T. Harrington, and C. H. Rameriz- Ronda Activity of N-formimidoyl-thienamycin and cephalosporins against isolates from nosocomially acquired bacteremia. Antimicrob. Agents Chemother. 21: Jacobs, R. F., G. L. Kearns, J. M. Trang, A. L. Brown, B. Marmer, J. C. McIntosh, F. L. Underwood, and R. B. Kluza Single-dose pharmacokinetics of imipenem in children. J. Pediatr. 105: Kahan, J. S., F. M. Kahan, R. Goegelman, S. A. Currie, M. Jackson, E. 0. Stapley, T. W. Miller, A. K. Miller, D. Hendlin, S. Mochales, S. Hernandez, H. B. Woodruff, and J. Birnbaum Thienamycin, a new P-lactam antibiotic. I. Discovery, taxonomy, isolation and physical properties. J. Antibiot. 32: Livingston, W. K., A. M. Elliott, and C. G. Cobbs In vitro ANTIMICROB. AGENTS CHEMOTHER. activity of N-formimidoyl thienamycin (MK0787) against resistant strains of Pseudomonas aeruginosa, Staphylococcus epidermidis, Serratia marcescens, and Enterococcus spp. Antimicrob. Agents Chemother. 19: Myers, C. M., and J. L. Blumer Determination of imipenem and cilastatin in serum by high-pressure liquid chromatography. Antimicrob. Agents Chemother. 26: Neu, H. C., and P. Labthavikul Comparative in vitro activity of N-formimidoyl thienamycin against gram-positive and gram-negative aerobic and anaerobic species and its,- lactamase stability. Antimicrob. Agents Chemother. 21: Patamasucon, P., and G. H. McCracken, Jr Pharmacokinetics and bacteriological efficacy of N-formimidoyl thienamycin in experimental Escherichia coli meningitis. Antimicrob. Agents Chemother. 21: Schaad, U. B., G. H. McCracken, Jr., C. A. Loock, and M. L. Thomas Pharmacokinetics and bacteriologic efficacy of moxalactam, cefotaxime, cefoperazone, and rocephin in experimental bacterial meningitis. J. Infect. Dis. 143: Skopnik, H., and G. Heimann Pharmacokinetics of antimicrobial drugs in the cerebrospinal fluid. Pediatr. Pharmacol. 3: Snedecor, G. W., and W. G. Cochran Statistical methods, p Iowa State University Press, Ames. 18. Verbist, L., and J. Verhaegen In vitro activity of N- formimidoyl thienamycin in comparison with cefotaxime, moxalactam, and ceftazidime. Antimicrob. Agents Chemother. 19: Downloaded from on April 28, 2018 by guest