Inflamed Meninges. 2.0 to 3.5 kg were anesthetized by a slow intrave- was suspended in phosphate-buffered saline (PBS),

Transcription

1 ANTIMCROBIL AaGNzs AND CHzmOTHRAPY, Dec. 1977, p Copyright American Society for Microbiology Vol. 12, No. 6 Printed in U.S.A. Blood, Brain, and Cerebrospinal Fluid Concentrations of Several Antibiotics in Rabbits with Intact and Inflamed Meninges THOMAS R. BEAM, JR.,* AnD JAMES C. ALLEN Division ofinfectious Diseases, State University ofnew York at Buffalo, Buffalo, New York Received for publication 10 May 1977 Because cerebrospinal fluid (CSF) antibiotic levels fail to predict either clinical success or relapse in the treatment of bacterial meningitis, we examined simultaneous antibiotic concentrations in the blood, brain, and CSF of control rabbits and of animals with experimental pneumococcal meningitis. Cefamandole pharmacokinetics were analyzed in detail and compared with those of cephalothin, ampicillin, penicillin G, and tobramycin. After 4 h of continuous intravenous infusion, cefamandole reached concentrations in both brain and CSF in excess of the minimal bactericidal concentration for the test organism and compared favorably with ampicillin and penicillin in achieving bacteriological cure. Cephalothin levels in the central nervous system remained undetectable in both control and infected animals during this time. Tobramycin concentrations were measurable in the CSF, but not in brain tissue in association with an inflammatory stimulus. In treatment of bacterial meningitis, the ments analyzing cefamandole, a new cephalosporin antibiotic. A comparison with several assessment of the potential therapeutic efficacy of an antibiotic has generally been made by other antibiotics of known or questioned value determination of drug concentrations in the for the treatment of meningitis is made. cerebrospinal fluid (CSF) after parenteral administration of the drug in the presence of MATERIALS AND METHODS inflamed meninges. The defined goal is to exceed the in vitro minimal inhibitory concentra- 2.0 to 3.5 kg were anesthetized by a slow intrave- Adult, male New Zealand white rabbits weighing tion or minimal bactericidal concentration nous infusion of 30 mg of Diabutal (sodium pentobarbital) per kg, administered via a peripheral ear (MBC) of the infecting organism (15, 24, 27, vein. A 30). However, several authors have noted that cisternal puncture was performed with a 20-gauge needle and a 1.0-ml tuberculin syringe; CSF antibiotic levels do not uniformly correlate 0.5 ml of CSF was withdrawn. If visible blood with either clinical cure or failure of response contamination resulted, the animal was not included in the study. A type III pneumococcus ob- (3, 19, 26, 30). It is known that 33 to 50% of CSF formation is extrachoroidal, the bulk of tained from a child with otitis media was passed this being derived from passage through the five times intraperitoneally in mice and five times interstitial substance of the brain (4, 5, 22). intracisternally in rabbits. After overnight growth Moellering and Fischer have documented the on 5% sheep blood agar (Baktron), this organism clinical significance of differences between ventricular and lumbar CSF levels (17), thus lend- was suspended in phosphate-buffered saline (PBS), adjusted to 1.7 x 109 to 4.0 x 109 colony-forming units per ml, and injected via the same needle used ing support to the importance of extrachoroidal for the cisternal tap. Gentle reaspiration and injection were performed to insure intracisternal inoc- CSF formation. Finally, histological examination has revealed microabscesses in the cerebral substance, documenting extension of men- was returned to its cage for 18 h. ulation. The needle was withdrawn, and the animal ingitis beyond the leptomeninges in cases that The animals were then treated for the designated have failed to respond to antibiotics (1, 23). time interval with cefamandole (kindly supplied by Because of all of these poorly understood features of meningitis and its therapy, we mea- 4D), ampicillin (Bristol Laboratories, Syracuse, Eli Lilly & Co., Indianapolis, Ind., lot no. CT N. sured siimultaneously the drug concentrations Y., lot no. J6M11), penicillin G (E. R. Squibb & Sons, Princeton, N.J., batch no J577), cephalothin (Eli Lilly & Co., lot no B), or in blood, brain, and CSF, utilizing several antibiotics and a rabbit model previously described tobramycin (Eli Lilly & Co., lot no. 9VA84A). An (20). Reported here are the results of experi- intravenous line was started by using a 23-gauge 710

2 VOL. 12, 1977 butterfly needle, and the drug was delivered by continuous infusion with a Harvard infusion-withdrawal pump (model 940, Harvard Apparatus, Millis, Mass.) via a peripheral ear vein. At 10 min before completion of the infusion, blood was obtained from the opposite ear and allowed to clot. The sample was centrifuged in a Serofuge II (Clay- Adams, Inc., Parsippany, N.J.), and the serum was aspirated and stored at -70 C. The animal was sacrificed by barbiturate overdose. A cisternal tap was performed prior to respiratory arrest. A portion of aspirated CSF was saved for immediate cell count; 0.1 to 0.2 ml was streaked onto a blood agar plate and cultured in a 37 C CO2 incubator overnight, and the balance of the sample was quick-frozen in a dry ice-acetone bath and stored at -70 C. A midline incision was then made over the skull, and the skin was retracted. Multiple burr holes were placed, and a skull flap was removed. The brain was withdrawn in toto, washed once in PBS to remove surface blood contamination, quick-frozen, and stored at -70 C. Antibiotic concentrations were not significantly affected by this means of storage for up to 3 weeks. The collection of samples uniformly required less than 10 min after barbiturate overdose. Control animals were infused with the designated antibiotic, and samples were obtained as described above. Rates of antibiotic infusion were chosen to produce those drug concentrations frequently obtained in serum of humans and considared within the therapeutic range. Antibiotic concentrations were determined by means of an agar-well diffusion technique. For determinations in serum, an overnight growth of Bacillus subtilis ATCC 6633 was adjusted to 0.7 x 107 to 2.1 x 107 colony-forming units per ml by nephelometry. A 0.5-ml amount of the suspension was added to 19.5 ml of tryptic soy agar (Difco Laboratories, Detroit, Mich.) and poured into a petri dish (100 by 15 mm; Falcon Plastics, Oxnard, Calif.). A ml amount of each serum sample was placed in duplicate into wells cut in the surface. Plates were incubated overnight at 37 C, and zones of inhibition were measured with a stereoscope (Bausch & Lomb, Inc., Rochester, N.Y.) fitted with a micrometer stage. All sera for a particular antibiotic were processed simultaneously. Three measurements of each zone diameter were made, and the mean was utilized for statistical analysis. Brain tissue was homogenized in Mueller-Hinton ANTIBIOTIC CONCENTRATIONS IN THE CNS 711 broth (Difco) at a ratio of 1:2 (wt/vol) with a Tri-R microhomogenizer (Tri-R Instruments, Rockville Center, N.Y.) with less than 0.16-mm clearance. The homogenate was centrifuged at 17,000 x g for 2 h in a refrigerated Sorvall RC2-B (Ivan Sorvall Inc., Norwalk, Conn.). The precipitate was found to have concentrations of antibiotic below the limits of detectability. The supernatant was passed through a 0.45-,Lm membrane filter (Millipore Corp., Bedford, Mass.) and used for antibiotic assay. CSF and brain tissue were corrected for contamination with blood by measurement of hemoglobin concentration after the method of Lowry and Hastings (14) and a calculation utilizing hematocrit. For the assay of brain and CSF antibiotic levels, a culture of B. subtilis was diluted 1:20, and the agar was added to a petri dish (150 by 15 mm; Fisher Scientific Co., Pittsburgh, Pa.). This modification increased the sensitivity of antibiotic detection to the following levels (in namograms per milliliter): cefamandole, 50; cephalothin, 200; ampicillin, 200; tobramycin, 50; and penicillin G, 50. CSF or brain values reported below these limits resulted when the following levels (in nanograms per milliliter): ing blood was subtracted from the experimentally determined concentrations. A standard curve was included on each plate. Two animals receiving no antibiotic infusion and not infected with pneumococci plus two animals infected but untreated were sacrificed; samples of serum, brain, and CSF were obtained as previously outlined. No inhibitory effect of any ofthese samples was documented by the agar-well diffusion technique utilized. Similarly, standards were prepared in serum, CSF, and brain homogenate from these animals and compared with standards made in PBS. No statistically significant differences were noted. Therefore, standards prepared in PBS were used for the assay technique. The validity of this procedure has been documented previously (24, 26). Standard curves were pooled because minimal variability was shown from plate to plate. Table 1 shows the consistency of the size ofzone of inhibition for cefamandole in the serum and central nervous system (CNS; CSF or brain). The zone of inhibition was then plotted against the logarithm of the antibiotic concentration, and the results were used to generate a computer program for the calculation of unknowns by the method of least squares. For ex- TABLz 1. Development ofcefamandole standard curves in serum and CNS from pooled data Cefaman- No. of dedole termina- Mean zone of +SDa concn tin inhibition conc tioi (diameter ml) in mm) Serum CNSb a SD, Standard deviation. b CSF or brain homogenate.

3 712 BEAM AND ALLEN ample, the summation plot of log10y = A x Bx, where x is the zone of inhibition in millimeters and Y is the concentration of antibiotic in micrograms per milliliter for cefamandole in CNS, had values of B = , A = , r = 0.987, and standard error of estimate = Individual values of B ranged from to 0.367, but 34 of 39 determinations were between and Similarly, values ofa ranged from to , but 33 of 39 determinations were between and The minimal inhibitory concentration and MBC of each antibiotic for the strain of pneumococcus used were determined in brain heart infusion broth (Difco) supplemented with nicotinamide adenine dinucleotide and heme (2). RESULTS It was decided to avoid the early peak in the blood antibiotic concentration produced by the administration of a loading dose. However, a steady-state level was attained for each antibiotic during h 1 of constant infusion. Within each treatment group there was a wide variation of serum antibiotic concentrations. For an individual animal, however, relatively constant levels were achieved throughout the course of the experiment. Thus, a representative animal from each group had hourly levels (in micrograms per milliliter) as follows: cefamandole, 40.9, 46.1, 47.4, and 48.4; cephalothin, 10.9, 12.3, 11.9, and 10.8; ampicillin, 11.9, 11.4, ANTIMICROB. AGENTS CHZMOTHZR. 9.5, and 8.6; penicillin G, 11.3, 12.5, 10.2, and 11.8; and tobramycin, 6.3, 9.2, 13.3, and 9.8. No threshold phenomenon for CNS penetration was documented with any of the antibiotics used by varying infusion rates up to twofold greater than the final rate chosen. The serum concentrations obtained are similar to those reported in humans (3, 6, 13, 18, 30). The pharmacokinetics of cefamandole penetration into CSF and brain in control animals are shown in Table 2. There was essentially complete exclusion of this antibiotic from the CSF throughout the experimental period. Brain penetration rose during the first 4 h of the infusion and remained relatively stable during the subsequent 4 h at a level times the serum concentration. Induction of an inflammatory response by intracisternal inoculation of pneumococci increased the absolute level of cefamandole in the CSF (Table 2). Similarly, a greater-than- 10-fold increase in the brain-to-serum ratio of cefamandole concentration was documented in h 1 of the infusion comparing intact and inflamed meninges. There was a progressive decline in the concentration of this antibiotic between h 2 and all subsequent hours of the experiments in both brain and CSF. The level of antibiotics in either CSF or brain substance did not correlate with the quantity of leuko- TABLz 2. Pharmacokinetics of cefamandole penetration of blood-brain and blood-csf barriers (40 mglkg per h) Duration Intact meninges Inflamed meninges ofif- No. of Srmlvl SFlvl Baneel No. of Seu ofn N eu level CSF level Brain levellev CSF level Brain level level (h) ama; (/Ag/Ml) (,ug/ml) (JAglg) mani- (;Lglml) (A/g) < < < < < <0.050 < < < < < < < < <

4 VOL. VANTIBIOTIC 12, 1977 CONCENTRATIONS IN THE CNS 713 cytes present at the termination of the experiment. Thus, 4 of 7 animals with CSF cefamandole levels exceeding 1.0 mg/ml had leukocyte counts less than 1,000 per mm3, whereas 7 of 10 animals with less than 0.35,ug/ml had leukocyte counts greater than 1,000 per mm3. Comparative data for 4-h infusions of all antibiotics studied in rabbits with noninflamed and inflamed meninges are shown in Table 3. In control animals, cephalothin did not penetrate either the blood-csf or the blood-brain barrier within the limits of detectability. In infected animals, hemoglobin contamination of the samples accounted for all antibiotic detected in two of three samples where levels were recorded; no cephalothin was detected in brain extracts. Ampicillin penetrated both barriers to a greater degree than did any other antibiotic tested. An inflammatory meningeal response increased CSF penetration, but had minimal effect upon brain penetration. Penicillin levels increased markedly in both CSF and brain in response to the pneumococcal inflammatory stimulus, whereas tobramycin was detectable in each CSF sample but in none of the brain homogenates. TABLE 3. All antibiotics, in which detectable levels were recorded, showed enhanced penetration into CSF in the experimental meningitis model. Cephalothin levels never exceeded the minimum of detectability at 0.2,ug/ml. A comparison of the ratio of CSF to serum antibiotic levels in inflamed versus intact meninges showed a 57-fold increase in cefamandole penetration, which exceeded that for ampicillin (9.7), penicillin G (28.6), and tobramycin (8.6). An analysis of the ratios of brain to serum antibiotic levels in inflamed versus intact meninges revealed minimal change for cefamandole (1.4), ampicillin (0.6), and tobramycin (2.0), with the greatest effect documented for penicillin G (10.7). CSF levels of cefamandole in excess of the MBC for the test organism ( utg/ml) were recorded for only 1 of 3 animals after 1 h of infusion, but were found in 10 of 12 animals with infusions of 2 h or longer. Brain levels in excess of the MBC were documented for all animals independent of the duration of infusion. CSF cultures were positive in six of nine animals infused for 3 h or less, but were negative in five of six animals infused 4 h or longer. The culture-positive animal in this latter group Blood-CSF and blood-brain barrier (4-h infusions) Intact meninges Inflamed meninges Antibiotic Infusion rate No of Seru evel Brai l INo. of Serum (mg/kg per h) ani level CSF ee ri Nove ofi level CS evel Brain level mals (p-giml) (gtg/ml) (p-gg) mals (p,g/ml) (-gijm') (Ogig) Cefamandole < < < < Cephalothin <0.200 < <0.035 < <0.200 < < <0.200 < <0.200 < <0.200 < <0.200 < <0.200 < <0.200 < <0.200 Ampicillin < Penicillin Tobramycin <0.050 < < <0.050 < < <0.050

5 714 BEAM AND ALLEN had both CSF (6.379,ug/ml) and brain (0.353,ug/ml) levels in excess of the MBC for the pneumococcus. Thus, 4 h of antibiotic infusion was chosen as a critical time for the assessment of bacteriological results. Table 4 is a comparison of cefamandole with other antibiotics in achieving bacteriological cure after 4-h infusions. Cefamandole compared favorably with ampicillin in this series (five of six animals cured after a 4-h infusion) and was superior to cephalothin (two of six cured). The culture-positive animals treated with ampicillin had both CSF (1.999,ug/ml) and brain (0.395,ug/ml) levels of antibiotic in excess of the MBC for the organism. Similarly, the therapeutic failure to penicillin had both CSF (0.167 pg/ ml) and brain (0.331 p,g/ml) levels in excess of the MBC. Among the culture-negative animals, there was only one bacteriological cure that dissociated brain and CSF levels. In this animal the brain level of cefamandole was 0.117,ug/ml (greater than the MBC), and the CSF level was 0.036,ug/ml (less than the MBC). TABLz 4. AntibiocMBC ANTIMICROB. AGENTS CHZMOTHZR. DISCUSSION Several authors (21, 26, 29) including ourselves have noted wide animal-to-animal variation of serum levels achieved by the administration of antibiotics on the basis of weight. However, our data document minimal variability for an individual animal during constant infusion. Pharnacokinetic analysis of cefamandole and all other antibiotics is presented as compartmental ratios to minimize such variability. Few previous workers have assessed brain penetration of antibiotics (11, 12, 24, 25, 33), and none has assessed it in relation to pharmacokinetic, pathophysiological, and therapeutic considerations. An analysis of the brain and CSF as anatomically separate compartments, as well as a new knowledge of pharmacokinetics, suggests that continuous intravenous infusion may offer an advantage in the penetration of these barriers (9). Whether therapeutic advantage may be gained in the clinical setting remains to be determined. Hemoglobin contamination can be assessed visually to levels of 0.1 g/100 ml. However, even small quantities of blood in CSF could account for a large proportion of an antibiotic detected in nanogram quantities. Thus, the technique of Lowry and Hastings (14) was used to supplement visual inspection of samples. With brain homogenates, a significant amount of blood contamination can be eliminated by exsanguinating the animal prior to removal of the brain. On the other hand, neither the assessment of the effects of this procedure on tissue levels nor the dynamics of altered membrane permeability under these circumstances has been undertaken. Thus, brains were removed as quickly as possible and washed to remove surface blood contamination, and correction for hemoglobin contamination of the resulting homogenate was made by nephelometry. In only 6 of 84 measurements of the brain or CSF level of an antibiotic did the calculated blood correction exceed the amount determined by the zone of inhibition. Five of the six disparate values differed by less than 0.05,ug/ml. Waterman et al. (32) have recently assessed the importance of time for an antibiotic to become bound in serum. These authors reported high concentrations of interstitial cefazolin during the equilibration phase between the intravascular and tissue spaces (32). The data in Table 2 support this concept for cefamandole penetration into brain. Thus, higher concentrations are recorded for h 1 of infusion (ratio of brain to serum, 0.09) than for h 2 (0.04) with intact meninges. Cefamandole does not penetrate the intact blood-csf barrier to an appreciable degree. The small amounts present in brain may be bound by tissue protein, thus explaining the failure of cefamandole to appear in the CSF. An alternative explanation would be that a longer infusion time is required for significant Bacteriological response after 4 h oftherapy Positive CSF cultures Negative CSF cultures Total no. Concn of antibiotic Concn of antibiotic Antibiotic o(/m) f an- No. of > MBC No. of > MBC animals animals CSF Brain CSF Brain Control Cefamandole Cephalothin Ampicillin Penicillin Tobramycin >

6 VOL. 12, 1977 CSF penetration to occur. The latter seems unlikely, however, because of known characteristics of cefamandole lipid solubility, serum protein binding, and ionization (34). With inflamed meninges the highest brain penetration occurred during the first 2 h. A progressive decline followed through 8 h of infusion. Whether some of this decline may be attributed to the passage of the drug into the extrachoroidal CSF, while free drug in the serum was reduced by protein binding, remains to be documented. The potential effects of diminishing inflammatory response on barrier permeability in cotjunction with bacteriological cure must also be considered. Also, the relative importance of bulk flow or active transport of this new cephalosporin out of CSF and back into the venous system, as previously observed with penicillin (31) and gentamicin (28), has yet to be determined. Both ampicillin and penicillin G penetrated the intact blood-csf barrier. With ampicillin also, there were high concentrations of drug in brain tissue. Cephalothin and tobramycin were excluded by both barriers within the limits of detectability. However, of greater significance, cephalothin also failed to penetrate these barriers in detectable amounts in the presence of meningitis. Our results thus lend support to the growing data regarding therapeutic failures with this antibiotic (8, 16). Tobramycin appeared in the CSF but not in the brain in this experimental model of pneumococcal meningitis. Kaiser and McGee (10) first noted therapeutic failures in patients receiving parenteral aminoglycosides. They documented lumbar, cisternal, and ventricular levels below the MBC of the infecting organism in four of five cases (10). Also, residual ventriculitis has been a complication related to inadequate antibiotic effect. Moellering and Fischer have observed the persistence of ventricular infection with Proteus morgagni while lumbar fluid was sterile. Levels in these compartments were 1.2 and 10.3,ug/ml, respectively, while the MBC for the infecting organism was 8,ug/ ml (17). Thus, current recommendations for the use of these antibiotics in meningitis due to susceptible organisms include intrathecal (7) or, preferably, intraventricular (10) supplementation of parenteral administration. Finally, webs have been noted at postmortem examination secondary to the infectious process. These may alter CSF dynamics, reducing choroidal flow to the ventricles. One possible explanation for these therapeutic failures may be the inability to penetrate brain tissue, as documented here for tobramycin. This could minimize the potential of antibiotic addition to ANTIBIOTIC CONCENTRATIONS IN THE CNS 715 CSF via extrachoroidal CSF production, particularly in the lateral ventricles. It may also account for the development of abscesses within the cerebral substance during persistent, inadequately treated meningitis (1, 23). In this model of pneumococcal meningitis, both ampicillin and penicillin reach levels in excess of the MBC for the test organism in brain and CSF. If these levels are valid criteria for assessing therapeutic efficacy, then cefamandole may be a potential alternative for the treatment of pneumococcal meningitis. Previous work supports CSF penetration by cefamandole in experimental meningitis due to Streptococcus pneumoniae (26) and Haemophilus influenzae (29). In data reported here a bacteriological cure was attained with 4 h of infusion in five of six animals for both cefamandole and ampicillin and in two of three animals for penicillin G. Cephalothin failed to sterilize the spinal fluid in four of six animals under similar circumstances. However, the data presented here do not allow a clear distinction between the importance of CSF or brain levels, since culture positivity persisted in a limited number of animals when both CSF and brain tissue levels exceeded the MBC. Thus, additional factors such as time and local host defense mechanisms may play an as yet undefined role. Extrapolation of these results to human meningitis should be done with care. The pathogenesis of the disease is different in humans. Fonnation and circulation of CSF differ between rabbits and humans (4). Antibiotic penetration of the blood-csf barrier was not correlated with leukocyte count here, nor has this been found in dogs (30), but Thrupp and associates did find such a correlation in humans (30). In this experimental meningitis model, however, cefamandole compared favorably with ampicillin and penicillin G in the penetration of both CSF and brain and resulted in a rapid bacteriological cure in a similar number of animals. Antibiotic penetration of the brain may be an additional relevant factor in determining therapeutic efficacy of antimicrobial agents in bacterial meningitis. ACKNOWLEDGMENT We thank Sylvia Domanski for her skilled technical assistance. LITERATURE CITED 1. Berman, P. H., and B. Q. Banker Neonatal meningitis-clinical and pathological study of 29 cases. Pediatrics 38: Bodine, J. A., L. J. Strausbaugh, and M. A. Sande Ampicillin and an ester in experimental Hemophilus influenzae meningitis. Clin. Pharmacol Ther. 20:

7 716 BEAM AND ALLEN ANTIMICROB. AGZNTS CHEMOTHER. 3. Brown, J. D., A. W. Mathies, Jr., D. Ivler, W. S. Warren, and J. M. Leedom Variable results of cephalothin therapy for meningococcal meningitis, p Antimicrob. Agents Chemother Cserr, H. F Physiology of the choroid plexus. Physiol. Rev. 51: Dobbing, J The blood-brain barrier. Physiol. Rev. 41: Dumoff-Stanley, E., H. F. Dowling, and L. K. Sweet The absorption into and distribution of penicillin in the cerebrospinal fluid. J. Clin. Invest. 25: Editorial Intrathecal antibiotics in purulent meningitis. Lancet ii: Fisher, L. S., A. W. Chow, T. T. Yoshikawa, and L. B. Guze Cephalothin and cephaloridine therapy for bacterial meningitis. Ann. Intern. Med. 82: Greenblatt, D. J., and J. Koch-Weser Clinical pharmacokinetics. N. Engl. J. Mod. 293: , Kaiser, A. B., and Z. A. McGee Aminoglycoside therapy of Gram-negative bacillary meningitis. N. Engl. J. Med. 293: Kramer, P. W., R. S. Griffith, and R. L. Campbell Antibiotic penetration of the brain. J. Neurosurg. 31: Lithander, A The passage of penicillins into the cerebrospinal fluid and brain in experimental meningitis-experimental investigations on rabbits. Postgrad. Med. J. Suppl. 40: Lockwood, W. R., and J. D. Bower Tobramycin and gentamicin concentrations in the serum of normal and anephric patients. Antimicrob. Agents Chemother. 3: Lowry, 0. H., and A. B. Hastings Histochemical changes associated with aging. J. Biol. Chem. 143: McCracken, G. H., and S. G. Mize A controlled study of intrathecal antibiotic therapy in Gram-negative enteric meningitis of infancy. J. Pediatr. 89: Mangi, R. J., R. S. Kundargi, R. Quintiliani, and V. T. Andriole Development of meningitis during cephalothin therapy. Ann. Intern. Med. 78: Moellering, R. C., and E. G. Fischer Relationship of intraventricular gentamicin levels to cure of meningitis. J. Pediatr. 81: Naumann, P., and E. Reintjiens Antibacterial activity and pharmacokinetic behavior of cefazolin as compared with five other cephalosporin antibiotics. Infection 2: Oppenheimer, S., H. N. Beaty, and R. G. Petersdorf Pathogenesis of meningitis. VIII. Cerebrospinal fluid and blood concentrations of methicillin, cephalothin, and cephaloridine in experimental pneumococcal meningitis. J. Lab. Clin. Med. 73: Petersdorf, R. G., and C. N. Luttrell Studies on the pathogenesis of meningitis. I. Intrathecal infection. J. Clin. Invest. 41: Plorde, J. J., M. Garcia, and R. G. Petersdorf Studies on the pathogenesis of meningitis. IV. Penicillin levels in the cerebrospinal fluid in experimental meningitis. J. Lab. Clin. Med. 64: Rail, D. P Drug entry into brain and cerebrospinal fluid, p In B. N. LaDu, H. G. Mandel, and E. L. Way (ed.), Fundamentals of drug metabolism and drug disposition. Williams and Wilkins, Baltimore. 23. Rorke, L. B., and F. W. Pitts Purulent meningitis-the pathologic basis of clinical manifestations. Clin. Pediatr. (Philadelphia) 2: Ruedy, J The concentrations of penicillins in the cerebrospinal fluid and brain of rabbits with experimental meningitis. Can. J. Physiol. Pharmacol. 43: Saha, A. L., and R. J. Joynt Meningitis, p In A. B. Baker and L. H. Baker (ed.), Clinical neurology. Harper and Row, Hagerstown, Md. 26. Sheretz, R. J., R. Dacey, and M. A. Sande Cefamandole in the therapy of experimental pneumococcal meningitis. J. Antimicrob. Agents Chemother. 2: Smith, D. H., D. L. Ingram, A. L. Smith, F. Gilles, and M. J. Bresnan Bacterial meningitis. Pediatrics 52: Spector, R The transport of gentamicin in the choroid plexus and cerebrospinal fluid. J. Pharmacol. Exp. Ther. 194: Strausbergh, L. J., C. D. Mandelaria, and M. A. Sande Cefamandole and ampicillin therapy in experimental Haenophilus influenzae meningitis. J. Infect. Dis. 135: Thrupp, L. D., J. M. Leedom, D. Ivier, P. F. Wehrle, B. Portnoy, and A. W. Mathies Ampicillin levels in the cerebrospinal fluid during treatment of bacterial meningitis, p Antimicrob. Agents Chemother Walters, I. N., P. F. Teychenne, L. E. Claveria, and D. B. Caine Penicillin transport from cerebrospinal fluid. Neurology 26: Waterman, N. G., M. J. Raff, L. Scharfenberger, and P. A. Barnwell Protein binding and concentrations of cephaloridine and cefazolin in serum and interstitial fluid of dogs. J. Infect. Dis. 113: Wellman, W. E., I. W. Dodge, F. R. Heilman, and M. C. Petersen Concentrations of antibiotics in the brain. J. Lab. Clin. Med. 43: Wick, W. E., and D. A. Preston Biological properties of three 3-heterocyclic-thiomethyl cephalosporin antibiotics. Antimicrob. Agents Chemother. 1: