Gentamicin Increases the Efficacy of Vancomycin against Penicillin-Resistant Pneumococci in the Rabbit Meningitis Model

In experimental meningitis a single dose of gentamicin (10 mg/kgof body weight) led to gentamicin levels in around cerebrospinalfluid (CSF) of 4 mg/liter for 4 h, decreasing slowly to 2 mg/liter4 h later. The CSF penetration of gentamicin ranged around 27%,calculated by comparison of areas under the curve (AUC in serum/AUCin CSF). Gentamicin monotherapy (-1.24 log₁₀ CFU/ml) was inferiorto vancomycin monotherapy (-2.54 log₁₀ CFU/ml) over 8 h againstpenicillin-resistant pneumococci. However, the combination ofvancomycin with gentamicin was significantly superior (-4.48log₁₀ CFU/ml) compared to either monotherapy alone. The synergisticactivity of vancomycin combined with gentamicin was also demonstratedin vitro in time-kill assays.

INTRODUCTION

Top

Introduction

The treatment of pneumococcal infections has been complicatedby the worldwide spread of penicillin-resistant strains (3).In life-threatening infections, particularly in meningitis,penicillin is ineffective even against strains with intermediateresistance, and penicillin resistance is often associated withresistance to other ß-lactam antibiotics. Becauseof treatment failures observed with cephalosporin monotherapy(4, 14), a combination of vancomycin with a broad-spectrum cephalosporin(ceftriaxone or cefotaxime) is usually recommended for treatmentof meningitis with resistant strains (12).

However, in case of ß-lactam allergy, the choice ofan adequate therapy is more challenging. Little is known aboutthe role of gentamicin in meningitis due to penicillin-resistantpneumococci. The aim of this study was to test gentamicin asmonotherapy and in combination with vancomycin against penicillin-resistantstrains in experimental meningitis. The standard treatment consistedof vancomycin combined with ceftriaxone.

MATERIALS AND METHODS

Top

Materials and Methods

Rabbit meningitis model. The meningitis model, originally described by Dacey and Sande(6), was slightly modified. The experimental protocol was acceptedby the local ethical committee (Veterinäramt des KantonsBern). Briefly, young New Zealand White rabbits weighing 2 to2.5 kg were anesthetized by intramuscular injections of ketamine(30 mg/kg of body weight) and xylazine (15 mg/kg), and wereimmobilized in stereotactic frames for induction of meningitisand cerebrospinal fluid (CSF) samplings. An inoculum containingapproximately 10⁶ CFU of penicillin-resistant pneumococci serotype6 was directly injected into the cisterna magna. The pneumococcalstrain had originally been isolated from a patient with pneumoniaat the University Hospital of Bern, Bern, Switzerland. The MICswere as follows (in milligrams per liter): penicillin, 4; ceftriaxone,0.5; vancomycin, 0.12 to 0.25; and gentamicin, 4.

A long-acting anesthetic (ethyl carbamate [urethane] 3.5 g/rabbit)was injected subcutaneously, and animals were returned to theircages. Fourteen hours later, a catheter was introduced intothe femoral artery for serum sampling, and the cisterna magnawas punctured again for periodic CSF sampling before and 1,2, 4, 6, and 8 h after initiation of therapy. Anesthesia wasperformed by repetitive intravenous (i.v.) injections of nembutal.Antibiotics were administered through a peripheral ear veinas bolus injections at the following concentrations: gentamicin,10 mg/kg; vancomycin, 20 mg/kg; and ceftriaxone, 125 mg/kg.Gentamicin and ceftriaxone were injected once at h 0, and vancomycinwas injected at h 0 and 4 according to the literature (1, 11).Untreated controls received saline. Bacterial titers were measuredby 10-fold serial dilutions of CSF samples, plated on bloodagar plates containing 5% sheep blood, and incubated overnightat 37°C. In parallel, 20 µl of undiluted CSF sampleswas plated (limit of detectability, 50 CFU/ml). Comparison betweendifferent dilutions of CSF was used to exclude significant carryovereffects during therapy. At the end of the experiment, euthanasiawas induced by a lethal i.v. dose of nembutal. The antimicrobialactivity of the regimens during the 8-h treatment was calculatedby linear regression analysis and expressed as a decrease oflog₁₀ CFU per milliliter per hour and as the killing rate after8 h. A value of 1.7 (log₁₀ of the limit of detectability) wasassigned to the first sterile CSF sample, and a value of 0 wasassigned to any subsequent sterile sample. The results wereexpressed as means ± standard deviation. Statisticalsignificance was determined by the Tukey multiple comparisonstest. All antibiotics and anesthetic drugs were commerciallypurchased.

Measurement of antibiotic levels in the CSF. Antibiotic concentrations in the CSF were determined by agardiffusion method. Standard curves were performed in saline with5% rabbit serum in order to mimic the protein concentrationin CSF during meningitis (5). Bacillus subtilis ATCC 6633 wasused as a test strain for vancomycin and gentamicin (7). Theintra- and interday variability of this method was less than10%. The limits of detection were 0.5 mg/liter for vancomycin,and 0.10 mg/liter for gentamicin.

In vitro assays. The pneumococcal strain was grown in C + Y medium (8) to anoptical density at 590 nm of 0.3 and then diluted 40-fold to10⁶ CFU/ml, corresponding to the CSF bacterial titer in rabbitsbefore initiation of therapy. Gentamicin was added in concentrationscorresponding to one time the MIC (4 mg/liter), correspondingto levels achieved in the CSF; vancomycin was added in concentrationsranging from one to two times the MIC. Combination therapy withvancomycin and gentamicin was also tested. Bacterial titerswere determined at h 0, 2, 4, 6, and 8 by serial dilution ofsamples; plated on agar plates containing 5% sheep blood; andincubated at 37°C for 24 h. Experiments were performed intriplicate, and results are expressed as means ± standarddeviation. Synergy was defined as bactericidal effect of a drugcombination greater than 2-log killing over the most activedrug alone.

RESULTS AND DISCUSSION

Top

Results and Discussion

We decided to administer gentamicin in a single dose in orderto produce higher levels in CSF than were achievable by conventionalmultiple daily dosing during the major part of the treatmentperiod (1). A single injection of gentamicin led to peak levelsin serum of around 33 mg/liter, declining to 2 mg/liter 8 hlater. The peak levels in CSF ranged around 4.3 mg/liter, andthe trough levels were around 2.2 mg/liter at the end of thetreatment period. The CSF penetration by gentamicin was calculatedfor each animal by comparison of serum and CSF areas under thecurve (AUC) (Systat software; SSPP Inc., Evanston, Ill.). Inour model, the penetration of gentamicin into the CSF was 27%± 7%, confirming previous studies (1). The CSF gentamicinlevels remained around the MIC (4 mg/liter) for approximately4 h (Fig. 1). The peak concentrations of gentamicin in serumwere similar to those observed in humans after one single-doseregimen (10). The CSF vancomycin levels ranged between 3.5 and1.5 mg/liter, remaining above the MIC during the entire treatmentperiod (data not shown), and corresponded to levels achievedin humans (2, 11).

View larger version (16K):
[in this window]
[in a new window]
FIG. 1. Gentamicin concentrations during 8 h in serum and CSF after a single dose of gentamicin (10 mg/kg). Open squares represent serum drug levels, and solid squares represent CSF drug levels. The concentration of gentamicin remained around the MIC (4 mg/liter) for approximately 4 h. Vertical bars represent ± standard deviation.

The killing rates of the different regimens are summarized inTable 1. Gentamicin monotherapy produced a negligible antibacterialactivity due to the pharmacokinetic profile of a single doseof gentamicin leading to CSF gentamicin levels around the MIConly during half of the treatment period (Fig. 1). In our experimentalmodel, vancomycin as either monotherapy or combined with ceftriaxoneshowed antibacterial activities comparable to those describedpreviously (8, 5, 13). It is interesting to note that the additionof vancomycin to gentamicin led to a synergy and significantlyincreased the killing rate of gentamicin, producing an antimicrobialactivity comparable to that of the standard regimen (vancomycinplus ceftriaxone). The synergistic activity between vancomycinand gentamicin was also found in vitro in time-kill assays over8 h. In this experimental setting, a gentamicin concentrationof one time the MIC (4 mg/liter) was selected, correspondingto levels achieved in the CSF of rabbits.

View this table:
[in this window]
[in a new window]
TABLE 1. Gentamicin, vancomycin, and combination therapy against penicillin-resistant S. pneumoniae in experimental meningitis^a

Addition of vancomycin in higher concentrations (two times theMIC) clearly resulted in synergy and sterilized the culturesafter 6 h (Fig. 2). Similar results were obtained with anotherpenicillin-resistant pneumococcal strain (KR4) in vitro (MICof penicillin, 4 mg/liter; MIC of gentamicin, 4 mg/liter) (datanot shown). Improved activity of gentamicin in vitro by additionof vancomycin has already been described, without reaching anextent qualifying as synergy (9). The reasons for the synergyobserved in vitro by increasing the vancomycin dose is not clear,but is reminiscent of the synergy between ß-lactamantibiotics as cell wall-active antibiotics and aminoglycosidesobserved in enterococci. A synergy between amoxicillin and gentamicinagainst resistant pneumococci has already been observed in amouse pneumonia model (77).

View larger version (22K):
[in this window]
[in a new window]
FIG. 2. Solid squares represent killing rates of vancomycin (Vanco2XMIC [0.25 mg/liter]), gentamicin (Genta1XMIC: [4 mg/liter]), and the combination of the two antibiotics (Genta1X MIC + Vanco 2XMIC) in vitro. Open squares represent untreated controls. The experiments were performed in triplicate, and the results are expressed as means ± standard deviations. *, P < 0.05 versus either single drug therapy.

Based on our data, it is obvious that gentamicin cannot be recommendedas monotherapy for pneumococcal meningitis due to resistantstrains because of its insufficient penetration into the CSFand its narrow safety profile. However, gentamicin combinedwith vancomycin could be a conceivable alternative regimen,especially in case of ß-lactam allergy. These datadeserve further evaluation.

FOOTNOTES

* Corresponding author. Mailing address: Department of Internal Medicine, Inselspital, 3010 Bern, Switzerland. Phone: 41316322111. Fax: 41316323847. E-mail: pcottagn{at}insel.ch.

REFERENCES

Top