J Int Med Res (1979) 7,45

Amikacin Concentration in the Cerebrospinal Fluid of Children with Acute Bacterial Meningitis Hugo Trujillo, MD, Heli Salgado, MD, Alvaro Uribe, MD, Nancy Agudelo, Carmen Tulia Zapata, Esther Luisa de Vidal and Ana Isabel Zuluaga, Antioquia University and Children's Hospital, Medellin, Colombia

Penetration of the aminoglycoside, amikacin, into the cerebrospinal fluid (CSF) of twenty children with acute bacterial meningitis was studied at various times after intramuscular administration and at various stages of therapy. Six of the patients were evaluated during therapy with amikacin at 7·5 mgfkg (intramuscularly) every 12 hours plus ampicillin every 6 hours at 300 mg/kg/day(intravenously); thirteen of the remaining fourteen patients were treated with ampicillin alone, but were given a single intramuscular dose of 7·5 mg/kg of amikacin for evaluation of CSF concentration. Amikacin concentration in CSF with respect to time after administration followed essentially the same pattern as in serum. A minimum concentration of 2 ugfml wasfound in 76% ofthe CSF samples obtained between 0·5 and 7 hours after administration. A mean amikacin serum/CSF ratio of 3:1 was demonstrated up to 7 hours after dose in all patients who underwent clinical improvement. Patient response was predictable by a correlation of in vitro MIC values with in vivo CSF concentration in three of the six patients who receivedamikacin therapy.

Introduction The use of a penicillin plus an aminoglycoside has long been accepted as initial therapy for neonatal bacterial meningitis (Le Frock, Prince & Klainer 1976, Lambert 1977, Yow et al 1973). More recently, this combination has also found favour as treatment for Gramnegative bacillary meningitis in adults (Guss & Wang 1977, Medical Letter 1977). Since success of therapy in patients of all ages has been predicated upon the ability of the antibiotics to penetrate the cerebrospinal fluid (CSF) in therapeutic concentrations, parenteral administration is often supplemented by intrathecal or intraventricular

administration of the amino glycoside (Kaiser 1975, Lee et al 1977). In the presence of ventriculitis, intrathecal supplementation is usually inadequate, while the neurosurgery required for intraventricular administration introduces more risk than is desirable for routine therapy. The solution to the problem does not depend solely on finding better ways to achieve higher antibiotic concentrations in the CSF. Indeed, after reviewing the lack of correlation between CSF antibiotic concentration and therapeutic outcome in neonatal meningitis, Lambert (1977) asked the question: 'Does the CSF concentration of & McGee

0300-0605/79/010045-07 $02·00

©Cambridge Medical Publications Limited

46

The Journal ofInternational Medical Research

antibiotic matter?' He concludes that, despite the uncertainties, the physician should aim for adequate antibiotic concentrations in the CSF and the blood, either by using a drug which achieves a high CSF level after systemic administration or by direct injection into the ventricles. The purpose of this paper is to present our findings on CSF penetration of the aminoglycoside, amikacin, when administered intramuscularly to children undergoing therapy for acute bacterial meningitis.

bacteria. Three enteric Gram-negative pathogens were characterized by the Gram stain method (Table 1). Six patients (Nos. 3, 8, 9, 10, 1I and 13) received amikacin as part of their therapeutic regimen. A clinical study of this group of patients has been published (Salgado, Trujillo & Uribe 1977). These six patients were treated for 2 I days with 7·5 mg/kg intramuscular amikacin every 12 hours plus intravenous ampicillin given every 6 hours at 300 mg/kg/day. Patients Nos. 3, 10 and I I also received an intrathecal injection of 5 mg amikacin each day for 3 days during this treatment period. Intrathecal administration was made at least 19 hours prior to evaluation and would not be expected to affect determinations of amikacin concentrations in CSF. In Medellin, Colombia, Haemophilus influenzae Type B is sensitive to ampicillin, thus, therapy for the four patients with meningitis caused by this pathogen, as well as

Materials and Methods Twenty children with acute bacterial meningitis were included in the pharmacokinetic study. The group consisted of thirteen males and seven females, ranging in age from 22 days to 10 years and in weight from 2·5 to 23 kg. Cerebrospinal fluid culture was used to identify Haemophilus influenzae, Streptococcus pneumoniae, Neisseria meningitidis and eleven Gram-negative enteric

Table 1 Characteristics, infecting pathogens and therapeutic outcome of paediatric patients with acute bacterial meningitis

Patient number

Age

Sex

1 2 3* 4 5 6 7 8* 9* 10" 11* 12 13* 14 15 16 I7 18 19 20

1 month II months 22 days 12 months 2 months 12 months 10 years 14 months 6+ months 3 months 2+ years 12 months I month 8 months 2 months 8 years 8 years 2 months 4 months 5 years

M M M M M M M F F F

M M M M F F

M M F F

Weight (kg)

Pathogen

Survive":

4·7 8·8 2·5 10·0 4·4 10·0 23·0 8·7 4·3 6·2 7·0 9·0 3·9 6·5 6·6 21·0 19·0 3·6 5-4 20·0

Escherichia coli Enteric Gram-negative Alcaligenes sp. Alcaligenes sp. Enteric Gram-negative Haemophilus influenzae Neisseria meningitidis Alcaligenes sp. Salmonella sp, Salmonella sp. Klebsiella sp. Haemophilus influenzae Proteus sp. Streptococcus pneumoniae Haemophilus influenzae Klebsiella sp. Salmonella sp. Enteric Gram-negative Haemophilus influenzae Streptococcus pneumoniae

Yes No No Yes No Yes Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes

"Patients who received amikacin as part of their therapeutic regimen. **All patients who survived were cured both clinically and bacteriologically.

Table 2

9

11

138

45

I 17

0·83 1·25

19**

22**

40

150

19

20

12·0

9·0

7·5

3·0

4·2

4·8

Lumbar Lumbar

46·6 25·0

18

24,000

253

Lumbar 27·6

Lumbar

315

Lumbar 38·4

64·0

2450 4700 Lumbar Lumbar 27·7 33·9

0

was not done

0

0

0

0 0

+++

53·0

3·0

33-0

47·0

13·0 22·0

35·5

61·0 6·0 47·0

6·0

9·0

3·7 59·0 0·0 7·5 12·0 6·0 26·0

8·0 0·0 0·0 61·0 0·0

mg/dl

Sugar

*Intrathecal amikacin was also administered to these 3 patients at 19 h. 24· 75 hand 22 h, prior to evaluation in patients Nos. 3, 10 and II, respectively. **The course of ampicillin therapy was completed in these 2 patients 4 days before amikacin was administered for pharmacokinetic evaluation.

1·00

12

4·7

5·00

27

5·0

13·0

1·17 1·3

5·0 3·5

18·0 10·3

1·00 2·00

18

--

10

8 8

30 150

13 16

404 Lumbar

29·1

3·5

12·0

1·00

6

50

15

--

0

+++

0

1000 I 150 1000

Lumbar Lumbar Lumbar

28·3 45·4 44·1

3·6 2·0 1·5

12· 7 4·4 3·4

0·77 5·00 6·00

50 30 50

14 13 II

+++

I 150

0 0

+++

0 0 40,000 0

0 was not done

2650 181 439 2100 6050 1900 5000

+++

0 0 0 0

Erythrocytes

3600

i

25,800 22,500 2100 66 154 I

No.lmm'

Leucocytes

Cerebrospinalfluid

Lumbar Lumbar Lumbar

49·1 28·2 42·8

6·0 2·4 1·2

12·2 8·5 2·8

4 4 4

3 3 3

45* 70 30

10 12 13

1·25 3·50 6·00

Lumbar Lumbar Lumbar Lumbar Lumbar Lumbar Lumbar

37·0 30·8 54·1 35·0 80·7 33·3 0·0

3·7 3·7 4·6 3·5 2·1 1·0 0·0

10·0 12·0 8·5 10·0 2·6 3·0 2·2

2 2 2 2 2 2 2

75 162 22* 65 32 45 50*

6 7 3 8 9 10 11

I

1 1 0·83 1·00 1·50 1·67 6·50 7·00 8·00

Ventricular Lumbar Lumbar Lumbar Subdural

CSF source

18·1 41·1 62·0 37·9 16·6

2·0 4·2 4·9 2·2 1·0

11·0 10·2 7·9 5·8 6·0

1 1

0·50 0·83 1·50 3·00 4·00

DaysJrom initiation oj treatment

37 60 22 75 35

Dosage (mg/dose)

1 2 3 4 5

Patient number

Interval between last Amikacin concentration intramuscular Serum CSF CSF dose and f-------evaluation (hours) (ugiml) %

Transfer of amikacin from serum to CSF at different stages of acute bacterial meningitis in 20 paediatric patients

56

3100

70

164

1830 67·6

214

23 795 132

795

2400

22·5 49 836 170 4800 480 130

2800 790 836 80 1270

Proteins

~

-.I

~

~

t0-

::;3

::t:

48

The Journal ofInternational Medical Research

for nine other patients consisted of ampicillin alone, given intravenously every 6 hours at 300 mg/kg/day. Patient No. 17 was treated with a combination of isoniazid and rifampin. In order to determine uptake of amikacin by CSF at various stages of the disease, these fourteen patients each received a single intramuscular injection of amikacin at 7·5 mg/kg on one predetermined day of their 10 to 21 day period of therapy with the other antibiotics. Cerebrospinal fluid samples were taken on different days during treatment, between 0·5 and 8 hours after intramuscular administration of amikacin. A portion of the CSF sample was used to determine cell, sugar and protein content and the remaining portion, plus a blood sample taken simultaneously, were used to measure amikacin concentration. A modification of the Simon and Yin method (1970) using Bacillus subtilis was used to determine amikacin concentration; ampicillin was neutralized with penicillinase (Waterworth 1973).

amikacin CSF/serum ratio as a function of post-administration time in patients with different degrees of meningeal inflammation and, therefore, does not readily lend itself to direct comparison of individual values. At all stages of therapy, the absolute concentration of amikacin in the CSF decreased with increasing time after administration, but this was a reflection of the changing serum concentration (Table 2) and bore no discernible relationship to the degree of meningeal inflammation. A concentration of at least 2,ug/ml was attained in nineteen of twenty-five (76%) of the CSF samples withdrawn between 0·5 and 7 hours after amikacin administration, independent of the day of therapy. In order to determine an overall pattern of amikacin penetration into CSF relative to the improving clinical status of the patients throughout the course of antibiotic therapy, the samples from the fifteen children who were cured were evaluated separately. When the averages of these values (Table 3) are plotted versus the day of therapy (Figure 1), it is evident that among the paediatric population who were cured, penetration into the CSF was maintained at a fairly constant level of about 33% during the first 12 days of therapy, as well as in two instances which were measured 4 days post-therapy. Patients Nos. 19 and 20 had been cured of acute bacterial meningitis by ampicillin therapy 4 days before receiving the injection of amikacin for the pharmacokinetic study. In the one case where a complete analysis was done, the transparency of the CSF and cytochemical results further testified to the normal condition of the patient. Nevertheless, as Table 3 and Figure 1 demonstrate, the concentration of amikacin in the CSF and per cent transfer from the serum in these two patients was similar to the values obtained in samples from other patients at various stages of the disease. The sensitivities of the pathogens isolated from the six patients treated with amikacin plus ampicillin are presented in Table 4. Among the four patients who were cured by this therapeutic regimen, the minimum inhibitory concentration determined in vitro (Table 4) was attained by amikacin in vivo (Table 2) in the serum of three patients (Nos. 8, 10 and 13) and in the CSF of two patients (Nos. 10 and 13).

Results All determinations of amikacin concentration obtained from twenty-six serum and CSF samples, as well as chemical and cellular characteristics of the CSF, are presented in Table 2. These determinations were made on different days after initiation of therapy, as well as at different times after administration of the intramuscular test dose of amikacin. In the CSF samples, pleocytosis ranged from a nearly normal value of 18 leucocytes to a maximum of 24,000 leucocytes/rnm ' in samples taken by lumbar puncture. One determination was made on the ventricular fluid and this contained 25,800 leucocytes/rnm '. In six samples, erythrocytes were abundant. Sugar concentrations ranged from zero to 61 mg/dl. Protein levels were between 23 and 3100 mg/dl among non-bloodcontaminated samples. Of the twenty-four instances in which CSF was completely analyzed, twenty-two signified the presence of marked inflammation of the meninges, even among the six blood-contaminated samples. In patients Nos. 4 and 20, a minimal amount of inflammation was evidenced by relatively low numbers of leucocytes and high values for CSF sugar. The study was designed to determine the

H Trujillo et al

49

Table 3 Uptake of amikacin in 20 samples taken from 15 children with acute bacterial meningitis Patient number

Day of therapy

Amikacin in CSF(/-lg/ml)

Averages in CSF(/-lg/ml)

Transfer ofamikacin from serum to CSF (%)

Average % transfer

2·1

18·1 37·9

28·0

27·2

4

1 1

2·0 2·2

6 7 8 10 11

2 2 2 2 2

3·7 3·7 3·5 1·0 0·0

2·4

37·0 30·8 35·0 33·3 0·0

10 12 13

3 3 3

6·0 2·4 1·2

3·2

49·1 28·2 42·8

40·0

11 13 14

4 4 4

1·5 2·0 3·6

2·4

44·1 45·4 28·3

39·3

15

6

3·5

3·5

29·1

29·1

13 16

8 8

5·0 3·5

4·3

27· 7 33·9

30·8

17

9

5·0

5·0

38·4

38·4

10

11

1·3

1·3

27·6

27·6

19 20

4 days post-therapy 4 days post-therapy

4·2 3·0

3·6

46·6 25·0

35·8

I

Total number of samples

=

20

32·9

2·9

Table 4 Sensitivity of pathogens isolated from patients treated with amikacin plus ampicillin Sensitivity Patient number

Pathogen

Amikacin MIC (uglml)

Ampicillin zone size (mm)

3 8* 9 10* 11* 13*

Alcaligenes sp. Alcaligenes sp. Salmonella sp. Salmonella sp. Klebsiella sp. Proteus sp.

20 5 1·25 1·25 5 1·25

6 6** 21 25 22 6

*Patients who were cured. **Sensitivity determination made in ,ug/ml for this one ampicillin sample only.

50

The Journal ofInternational Medical Research

Discussion It has been generally accepted that aminoglycosides penetrate intact meninges poorly and, even when the meninges are inflamed, penetration is not increased significantly if the amino glycoside is administered only by the systemic route. Our study showed, within limits of variation anticipated when comparing samples taken from different patients, CSF concentrations of amikacin generally followed a pattern similar to those in serum. The highest CSF values were obtained within the first 2 hours after administration, they diminished during the next 5 hours, and by 8 hours after administration, amikacin was not detectable in the one CSF sample that was evaluated. Of the twenty-five samples withdrawn between 0·5 and 7 hours after an

intramuscular dose of 7·5 mg/kg of amikacin, nineteen (76%) samples exhibited concentrations of at least 2.ug/ml. These nineteen samples included one sample from each of two patients who were cured of meningitis following ampicillin therapy. While these two patients demonstrated a greater degree of penetration than twenty-four normal (with respect to meningeal irritation) subjects evaluated by Robson and Briedis (1977), the two studies may not be directly comparable due to differences in post-administration sampling times and the recently recovered state of the patients in our study. Our attempt to demonstrate a relationship between amikacin penetration and degree of meningeal inflammation showed that the average transfer of intramuscularly administered amikacin from serum to CSF did

N LL (f)

l!)

C')

C')

II

II

II

II

-S

-S

-S

-S

o

0

N

II c

II c

II c

u

.9 50 E 2Q) en

E 40

2

o

c

o

'(3

OJ

.:,£

'EOJ

30

o

'0

o

0

o

o

c

.g ...e

20

Q)

c

Q)

...c.c

~ 10

Ql

0..

2

4

6

8

10

12

I

I

(19 and 22) 4 days post-therapy

Day of therapy

Fig 1 Uptake of amikacin by 20 samples of CSF taken from 15 children during cure of acute bacterial meningitis (n = number of samples)

H Trujillo et al not decrease as the health of the patients improved but was maintained at approximately the same level throughout the first 12 days of therapy and even after cure had been effected. A constancy of aminoglycoside penetration independent of the degree of meningeal inflammation has also been reported following intrathecal administration of gentamicin (Rahal et aI1974). Since the pharmacological evaluation of amikacin was carried out concurrently with its use as therapy in six patients in our study, we examined the relationship of the concentration of amikacin in CSP, its MIC against the isolated pathogen and therapeutic outcome in each of these patients. Among the four patients cured, two had CSF concentrations lower than the MIC, and of the two patients who died, one had a CSF concentration greater than the MIC. Lack of correlation between CSF concentration and patient response is thus evident in our investigation just as it was in those studies which caused Lambert (1977) to question the importance of CSF antibiotic concentration. It is unlikely that any single approach wiil suffice to solve the many problems involved in the management of Gram-negative bacterial meningitis. Even when therapeutic success has been demonstrated there is often little agreement 01} the pharmacological basis for cure. Our study, while unable to provide any more difinitive answers to these questions, has presented heretofore unpublished information on the pharmacology of amikacin in acute bacterial meningitis. Scattered reports on the therapeutic effectiveness of this new

51 aminoglycoside in Gram-negative bacterial meningitis are encouraging and further trials are warranted, particularly in the treatment of meningitis caused by organisms resistant to other aminoglycosides. REFERENCES Guss D & Wang R I H (1977) Therapy of Gram-negative meningitis. Drug Therapy (Hospital) Dec, 38 Kaiser A B & McGee Z A (1975) Aminoglycoside therapy of Gram-negative bacillary meningitis. New England Journal of Medicine 293,1215 Lambert H P (1977) Problems with neonatal meningitis. Journal of Antimicrobial Chemotherapy 3, 381 Lee E L, Robinson M J, Thong M L, Punthucheary S D, OngTH & NgK K (1977) Intraventricular chemotherapy in neonatal meningitis. The Journal ofPaediatrics 91, 991 Le Frock J L, Prince R A & Klainer A S (1976) Bacterial meningitis. The Journal ofthe American Pharmaceutical Association 16,208 Medical Letter (1977) Intralumbar and intraventricular therapy of bacterial meningitis. Medical Letter 19,94 Rahal J J Jr, Hyams P J, SimberkotTM S & Rubinstein E (1974) Combined intrathecal and intramuscular gentamicin for Gram-negative meningitis. New England Journal ofMedicine 290, 1394 Robson H G & Briedis D J (1977) Cerebrospinal fluid penetration of amikacin in the absence of meningeal irritation. Clinical Research 25, 699A Salgodo H, Trujillo H & Uribe A (1977) Amikacin (BBK8) and ampicillin in acute bacterial meningitis of children. (Spanish) Investigacion Medica Internacional4, 300 Simon H J & Yin E J (1970) Microbioassay of antimicrobial agents. Applied Microbiology 19,573 Waterworth P M (1973) An enzyme preparation inactivating all penicillins and cephalosporins. The Journal of Clinical Pathology 26,596 Yow M D, Baker C J, Barrett F F & Ortigoza CO (1973) Initial antibiotic management of bacterial meningitis. Medicine 52,305