T HIS paper is concerned with antibiotic

!. Neurosurg. / Volume 31 / September, 1969 Antibiotic Penetration of the A Comparative Study PAUL W. KAME, M.D.,* ICHAD S. GIFFITH, M.D.,t AND OBET L. CAMPBELL, M.D. Neurological Surgery, Indiana University Medical Center, Indianapolis, Indiana, and Lilly Laboratories for Clinical esearch, Marion County General Hospital, Indianapolis, Indiana T HIS paper is concerned with antibiotic penetration of the brain in patients having non-inflamed meninges. It establishes the amount of antibiotic penetration of actual extravascular cerebral tissue. Such information may be useful in the antibiotic prophylaxis of basilar fractures, prophylaxis in routine craniotomy, or in the management of open cranial-cerebral wounds. Such data might also be pertinent to the therapy of established brain parenchymal infection in which the meninges are not necessarily involved, in particular, brain abscesses. There are many studies in the literature on cerebral spinal fluid levels of antibiotics in the non-infected central nervous system following systemic injection. Most antibiotics are recovered in the cerebrospinal fluid in little, or very small, amounts after the usual, or even increased, systemic doses. In the most carefully done studies, it has been concluded that the blood-brain barrier and cerebrospinal fluid barrier are not the same. 2 It can then be expected that actual drug penetration of the cerebrospinal fluid would differ from drug penetration of the extravascular cerebral tissue. With particular reference to antibiotics, it should be noted that the single other study on antibiotic penetration of cerebral tissue, which was conducted in 1954, attempted to correlate brain levels and cerebrospinal fluid levels and found this discrepancy in penetration of the two media. 1~ Five antibiotics were chosen for this study: 1) chloramphenicol, 2) cephalothin, 3) ampicillin, 4) penicillin G, and 5) cephaloridine. Chloramphenicol is a well-established drug and penetrates many tissues with eceived for publication October 14, 196. * Present address: 4640 N. Federal Highway, Fort Lauderdale, Florida 3330 -~ Senior Clinical Pharmacologist, Lilly Laboratories for Clinical esearch. 295 ease. Therapeutic levels have been recovered from the cerebrospinal fluid of patients with non-inflamed meninges, lz Spinal fluid levels have approximated one-fourth of the blood levels. 11,~ Cephalothin, a newer broad spectrum antibiotic, has had particular success against penicillin-resistant staphylococci. Its use, to date, in the management of central nervous system inflammatory disease has been limited. In non-inflamed meninges, it diffuses poorly into the cerebrospinal fluid2 Ampicillin, also a newer broad spectrum antibiotic, was early recognized as an effective drug in the management of meningitis? Since then, it has become the antibiotic of choice in many centers for the initial therapy of meningitis. It readily passes into the cerebrospinal fluid when the meninges are inflamed. 9 However, like cephalothin, the amount found in the CSF in normal individuals is insignificant. ~ Penicillin a is a well established antibiotic in the management of infection due to penicillin-sensitive, grampositive pathogens. Cephaloridine is a recently released, broad spectrum, bactericidal antibiotic chemically related to cephalothin. Materials and Methods Human brain tissue was analyzed for the five antibiotics. In all cases, the analyzed tissue had been removed as part of the usual surgical treatment of the basic disease or lesion. By necessity, then, most cases selected were either those with brain tumor or those requiring temporal lobectomy following trauma. In each case, effort was made to select as near normal tissue as possible, away from the primary lesion. Only one antibiotic could be given each patient and this was administered prior to surgery. The dosage of four antibiotics was the same, namely, 2 gm. The dose of penicillin Q calculated to be equal to 2 gm was 3.2 million units. Chlor-

296 P.W. Kramer,. S. Griffith and. L. Campbell FtG. 1. concentrations of chloramphenicol. amphenicol was given intramuscularly, while ampicillin, cephalothin, penicillin Q, and cephaloridine were given intravenously. The times of drug administration and specimens obtained were noted. Paired blood samples were taken at the time of each specimen, when possible. Aseptic handling of tissue was required as antibiotic levels were determined by biological assay. The brain samples varied from 50 to 300 mg. Each was frozen, ground into an emulsion, and a known amount of buffer solution added. Assay for all five antibiotics was by the modified cup-plate method using Sarcina lutea as the test organism. 4 Since the primary interest was in the actual amount of anti- FIG. 2. concentrations of chloramphenicol before and after correction for blood-antibiotic contamination. biotic in brain tissue itself, and since the specimens were necessarily contaminated with their own blood volume, as well as free blood, each sample was corrected for blood content. This was accomplished by spectrophotometric analysis of the homogenates for their hemoglobin content. 7 The blood-antibiotic contamination could then be calculated and subtracted from the specimen concentrations to obtain the amount of antibiotic per gram of extravascular brain tissue. esults Measurable antibiotic levels were found in the brain for all five antibiotics assayed. Corrections for blood contamination in the Fro. 3. concentrations of cephalothin.

Antibiotic Penetration of the 297 FIG. 4. concentrations of cephalothin before and after correction for blood-antibiotic contamination. samples varied and were usually small. Figure 1 shows the data for chloramphenicol. The blood levels ranged from 3.2 to 6.2 /~g/cc, with an average of 4 ~g. Measurable brain tissue levels were obtained in all samples assayed for chloramphenicol (Fig. 2). The range was from 9.7 to 412 /zg/cc following correction for blood in the samples. There were two inconsistent figures in the 14 brain assays shown on the graph. These were not included in the calculations of average brain content. The average chloramphenicol concentration was 36.0 ~g/gm of brain tissue. Figure 3 shows the blood levels after cephalothin administration. These ranged from 0.5 to 4 /~g/cc, with an average of 11.7 ~g. Measurable brain levels following correction were obtained in all samples (Fig. 4). These ranged from 0.4 to 2 ~g/grn of brain tissue. A single assay of 12 is inconsistent and is omitted from further calculations. The average corrected concentration was 1.6 /zg/gm of brain tissue. Figure 5 shows the blood levels following administration of ampicillin. These ranged from 5. to 45 ~g/cc, with an average of 21.0 ~g. concentrations following correction for blood contamination could not be found in four of the specimens obtained within 2 hours after antibiotic administration. However, brain tissue concentrations were found regularly after 2 hours. The highest corrected concentration was 2 /xg/gm of brain tissue (Fig. 6). The average was 0.4/~g/gm. Fir;. 5. concentrations of ampicillin.

29 P.W. Kramer,. S. Griffith and. L. Campbell FIG. 6. concentrations of ampicillin before and after correction for blood-antibiotic contamination. Arrows indicate values of zero antibiotic in brain after correction. Figure 7 shows the blood concentrations of penicillin G. These ranged from 2.20 to 17 ~u.g/cc, with an average of 7.50?.g. In two specimens a corrected brain tissue concentration could not be found. However, antibiotic was found in all other specimens, and the highest corrected concentration was 1.3 ~g/grn. The average concentration was 0.32.ag/gm (Fig. ). Figure 9 shows the blood concentrations of cephaloridine which ranged from.35 to 26.20 ~tg/cc, the later figure being a 45- minute sample. The average blood concentration was 1.05 ~g/cc. tissue concentrations following correction for blood contamination could not be found in four specimens. However, antibiotic was found in most specimens and the highest correct concentration was 1.97 ~.g/gm. The average concentration was 0.90 o.g/gm (Fig. 10). Table 1 demonstrates the numerical data from the previous figures. To make each of the five series as comparable as possible, specimens obtained before 1 hour and later than 4 hours have been omitted. We have also eliminated assays which were grossly inconsistent in two cases, as noted. These assays were obtained from the smallest brain samples, necessarily leading to the largest possible calculation errors. Average values FIo. 7. concentrations of penicillin '~.

Antibiotic Penetration of the 299 Fie.. concentrations of penicillin ~ before and after correction for blood-antibiotic contamination. Arrows indicate values of zero antibiotic in brain after correction. of blood and brain concentrations are also shown in Table 1. Table 2 summarizes the average blood levels and brain tissue concentrations for each antibiotic. From these data, a ratio can be calculated comparing the micrograms of antibiotic per cubic centimeter of blood with the micrograms of antibiotic in the comparable amount of brain, the gram. The values for chloramphenicol are striking. Thus, for 1 g.g of chloramphenicol per 1 cc of blood, there are 9 ~g of antibiotic in 1 gm of brain tissue. The figures for the other four antibiotics are more nearly that which would be expected. Thus, for cephalothin, there was only 1 ~g of antibiotic in 1 gm of brain tissue for each 7 ~tg/cc of the patient's blood. The blood to brain ratio for ampicillin was 56: 1, for penicillin 23:1, and for cephaloridine 20:1. To ascertain the effect of steroids on antibiotic penetration of the brain, the figures for chloramphenicol were used. In this group, five subjects received the usual neurosurgical doses of steroids (dexamethasone), and six cases were controls. The results are summarized in Table 3 and show remarkably similar brain-antibiotic levels. In each of the other antibiotic series either the number receiving or not receiving steroids was too small to give meaningful data. Discussion There are doubtless many variables which might influence the penetration of any drug TABLE 1 and corrected brain concentrations of antibiotics Chloramphenicol Cephalothin Ampicillin Penicillin Cephaloridine ug/cc ug/gm ug/cc ug/gm ug/cc ug/gm B ood ~1 ;/cc ag/gm ~g/cc,g/gin -- 47.0 -- 6.2 54.0 19.2 3.5 4.6 1.0 25.1 3.4 24.0 3.2 10.0 -- 30.0 3.4 "412.0 -- 63.0 4.0 36.0 46. 13.4 1. 27.0 1.4 2.1.7 11.7.4 1.7 3.0. *2 1.7 2.7 --72.6- h 3.0 0 27.0 0 45.0 0 10.0 2.0 24.0 0 19.0.3 15.0.3 6.0.5 21.0 0.4 15.0 0 5.0 0.10 1~.6 1.3 2.2 0.9 0.07 ~.9 0.66 5.1 0.02 2.3 0.07 0.56 7.5 0.32 15. 1.97 19.4 0 22.6 0.57 22. 1.5 20.4 0 16. 1.6.5 0.6 1.O5 O.9 * Not included in calculation.

300 P.W. Kramer,. S. Griffith and. L. Campbell Fro. 9. concentrations of cephaloridine. into the brain. Table 3 analyzes one variable, namely, steroid effect. Since one of the proposed mechanisms of steroid effect in cerebral edema is maintenance of the integrity of the blood-brain barrier, it might be expected to have some effect on drug penetration. Chloramphenicol penetration was not affected by steroid therapy. Antibiotics are well known to have differing effects on different bacterial species. To illustrate this, and to translate these data to the more meaningful clinical situation, several organisms have been analyzed. Table 4 lists the organisms that comprise the majority of those reported as the etiological agents in brain abscesses and craniotomy infections. Hemophilus has been included, because it easily completes the list of the three most common agents causing meningitis, although this is not the topic of this study. This table TABLE 2 -brain ratio using corrected brain levels between 1 and 4 hours shows the average in vitro concentration required of the five antibiotics to inhibit each of these organisms2, The variability of potency of each antibiotic against a given organism is marked in most cases. For example, the resistant staphylococcus is inhibited by only ~g of cephalothin, while 4.0 ~g of chloramphenicol (20 times as much) are required for bacterial inhibition. This same organism is resistant to ampicillin and penicillin ~ The number of micrograms of drug available in the brain assays to that required to inhibit these organisms may be compared. This establishes a comparative antibacterial activity for each of the five antibiotics in brain. Thus, Table 5 shows a multiple amount of antibiotic found in the brain specimens, as compared to that required to inhibit a given organism. For example, there is eight times the usual inhibitory dose of cephalothin or chloramphenicol for the staphylococcus in our brain specimens. It is noteworthy, also, that there is more than an adequate amount of chloramphenicol present in brain for inhibition of all of these organisms except the Antibiotic Chloramphenicol 7~ephalothin Ampicillin Penicillin ~ephaloridine Average Levels (ug/cc) 4.0 11.7 21.3 7.5 1.04 Average Levels (~,g/gm) 36.0 1.6 0.4 0.32 0.90 atio 1/9 7/1 56/1 23/1 20/1 TABLE 3 Effect of steroids on chloramphenicol penetration of the brain No. Cases Avg. Levels Without steroids 6 41 ~g/gm With steroids 5 40 ~g/gm

Antibiotic Penetration of the 301 Fro. 10. concentrations of cephaloridine before and after correction for blood-antibiotic contamination. Arrows indicate values of zero antibiotic in brain after correction. pseudomonas. Adequate levels of the other four antibiotics were present for the inhibition of sensitive gram positive cocci. Although corrected brain levels represented only very small amounts of drug for these four, 1.60 ~g/grn for cephalothin and less than 1 ~g/gm for the other three, it should be noted that the inhibitory concentrations required for these cocci were exceeded many times. The largest value was found for the streptococcus. Over 100 times the required level of penicillin and 300 times the required level of cephaloridine were present. None of the drugs tested, except chloramphenicol, had sufficient brain-antibiotic concentrations to inhibit the gram-negative organisms listed. Summary 1. Five antibiotics were studied for penetra- tion of the non-inflamed human brain. 2. A method for correcting brain tissue concentrations for blood-antibiotic contamination was utilized. 3. All five antibiotics, chloramphenicol, cephalothin, ampicillin, penicillin ~ and cephaloridine, gave measurable concentrations in the brain. However, they varied widely in actual tissue concentrations and in the ratio of blood-to-brain antibiotic. 4. All five antibiotics reached levels sufficient to inhibit sensitive gram-positive cocci. Penicillin-resistant staphylococci were not inhibited by concentrations of ampicillin and penicillin a obtained from brain tissue. 5. Chloramphenicol had the widest spectrum of activity because it achieved blood and brain tissue concentrations greater TABLE 4 Concentrations of antibiotics required to inhibit common intracranial pathogens Chloramphenicol Cephalothin Ampicillin Penicillin Cephaloridine Pathogen ug/cc ug/cc ug/cc #g/cc ug/cc esistant staphylococcus Staphylococcus Streptococcus Pneumococcus Proteus E. coli Klebsiella Aerobacter Pseudomonas H. influenzae 4.0 4.0 2.0 2.0 15.0 6.0 3.0 12.0 0.5 0.1 2.0 12.0 25.0 6.25 * 0.1 0.02 5.0 5.0 0.5 0.003 0.006 2.5 0.003 0.01 12.5 6.2 1.6 6.25 * = resistant.

302 P. W. Kramer,. S. Griffith and. L. Campbell TABLE 5 Comparative antibacterial activity of antibiotics in brain. (These figures represent a multiple of the amount of antibiotic found in brain samples as compared to that required to inhibit each organism.) Pathogen ~ Chloramphenicol Cephalothin Ampicillin Penicillin Cephaloridine esistant staphylococcus Staphylococcus Streptococcus Pneumococcus Proteus E. coli Klebsiella Aerobacter Pseudomonas H. influenzae 17 17 2 6 11 3 66 16 2 4 19 1 16 107 53 4.5 45 3O0 90 * Concentration in brain tissue less than the sensitivity ol the organism. than that required to inhibit any of the gram-negative organisms tested, except the pseudomonas. 6. The amount of chloramphenicol assayed from brain was nine times greater than the comparable blood level in a paired sample. This ratio was reversed for the other four drugs. 7. Steroid therapy did not affect brain penetration of chloramphenicol. Acknowledgments The authors are gratefully indebted to obert Wolen, Ph.D., and Mrs. Marjorie Joiner for their competent assistance in the preparation of this paper. eferences I. BONN, P., O'BIEN, J., BENTLEY, D., and HAYMAN, H. Further experience with ampicillin. Anti. microb. Ag. Chemother., 1962, 323-333. 2. DAVSON, H., and SMITH, H. V. Discussion on the penetration of drugs into the cerebrospinal fluid. Proc. r. Soc. Med., 1957, 50: 963-966. 3. GAOD, L. P., and SCOWEN, E. F. The principles of therapeutic use of antibiotics. Br. reed. Bull., 1960, 16:23-2. 4. GOVE, D. C., and ANDALL, W. A. Assay methods of antibiotics. A laboratory manual. New York. Antibiotics Monographs No. 2. Medical Encyclopedia, 1955, 14-16. 5. IVLE, D., THUPP, L. D., LEEDOM, J. M., WEnLE, P. F., and POTI~OY, B. Ampicillin in the treatment of acute bacterial meningitis. Anti. microb. Ag. Chemother., 1963, 335-345. 6. LEE, C. C., and ANDESON,.C. and tissue distribution of cephalothin. Anti. microb. Ag. Chemother., 1962, 695-701. 7. LOWY, O. H., and HASTINGS, A. B. Histochemical changes associated with aging. I. Biol. Chem., 1942, 143:257-269.. NAUMANN, P. Bacteriological and pharmacological properties of cephalothin and cephaloridine. Postgrad. med. l., 1967, 43:26-31. 9. THUPP, L. D., LEEDOM, J. M., IVLE, D., WEHLE, P. F., POTNOY, B., and MATHIES, A. W. Ampicillin levels in the cerebrospinal fluid during treatment of bacterial meningitis. Anti. microb. Ag. Chemother., 1965, 206-213. 10. WELLMAN, W. E., DODGE, H. W., HEILMAN, F.., and PETESEN, i.c. Concentration of antibiotics in the brain. J. Lab. clin. Med., 1954, 43:275-279. 11. WILLIAMS, J. B., and DAT,.M. Chloramphenicol (Chloromycetin) concentration in cerebrospinal ascetic and placental fluids. Boston reed. Q., 1950, 1:7-10. 12. WOODWAD, T. E., and WISSEMAN, C. L. Chloromycetin (Chloramphenicol). New York. Antibiotics Monographs, Medical Encyclopedia, 195, 13-1.