Journal of Antimicrobial Chemotherapy (1992) 29, 427^*33

Comparison of gentamidn dosing regimens using an in-vitro model

E. J. Begg*, B. A. Peddle', S. T. Chambers* and D. R. BoswdT

Departments of "Clinical Pharmacology, ^Infectious Disease, 'Pathology,

''Microbiology, Christchurch Hospital, Christchurch, New Zealand

An in-vitro model which simulates in-vivo pharmacokinetics was used to compare

the efficacy against Pseudomonas aeruginosa of dosing regimens of gentamicin which
achieve different peak/trough concentrations but use the same total dose over 24 h.
First exposure to gentamicin produced a rapid bactericidal effect which was propor-
tional to the initial peak concentration. Subsequent doses of gentamicin produced a
smaller bactericidal effect Regrowth occurred with all dosing regimens, even after
very high initial concentrations (26 mg/L). The time to reach bacterial counts above
starting values was prolonged in relation to peak concentrations. Regrowth was also
demonstrated in continuous infusion experiments which maintained very high
concentrations (26 mg/L), although an inhibitory effect was evident compared with
single dose experiments and the experiments mimicking in-vitro pharmacokinetics.
There was little evidence of a post-antibiotic effect The data supports the use of
larger initial and longer interval bolus dosing compared with current
recommendations.

Introduction
The aminoglycoside antibiotics have a low therapeutic index, and the choice of dose
regimen is critical. The importance of adequate peak concentrations for good clinical
outcome is now accepted (Noone et al., 1974; Moore, Smith & Lietman, 1984; Moore,
Leitman & Smith, 1987) but high doses carry the risk of nephrotoxicity and ototoxicity.
Peak concentrations of 6-10 mg/L and trough concentrations of < 2 mg/L are widely
used target concentrations which provide reasonable efficacy with low toxicity. Recent
evidence suggests that efficacy may be improved without increased toxicity by using
larger doses less frequently, to give higher peak and lower trough concentrations
(Chan, 1989). The reason for this may relate to the findings that aminoglycosides have
significant post-antibiotic effect (suppression of bacterial growth which persists after
short exposure to suprainhibitory concentrations of antibiotic) (Vogelman & Craig,
1985), concentration-dependent bactericidal activity (Moore et al., 1987), and toxicity
which may be diminished with longer dosing intervals (Powell et al., 1983).
The aim of this study was to compare the bacterial kill and regrowth patterns during
four dosing regimens of gentamicin which differ in peak and trough concentrations and
dosing interval but which have the same total dosage over time (equal area under the
aminoglycoside concentration time curves (AUQ), using a dynamic in-vitro model
which mimics in-vivo pharmacokinetics.
427
03O5-74S3/92/O4O427+07 $02.00/0 © 1992 The Britiih Society for Antimicrobial Chemotherapy

Downloaded from https://academic.oup.com/jac/article-abstract/29/4/427/825215

by guest
on 19 March 2018
 428

oluent
magnetic •ample
•tirrer
Figure 1. Diagrammatic representation of the equipment used in aL experiments.

Methods
Pseudomonas aeruginosa NCTC 10662 (gentamicin MIC of 2 mg/L) in the logarithmic
phase of growth was inoculated into a chamber containing Mueller-Hinton broth
supplemented by 12-5 mg/L calcium chloride and 12-5 mg/L magnesium chloride
(Figure 1). The culture was agitated by a magnetic stirrer and the temperature
maintained at 37CC by a water bath. Gentamicin was added to the reservoir and a
computer controlled syringe pump introduced timed additions of broth to produce an
exponential decline in gentamicin concentrations with a rate constant (k) of 0-2772
(based on a half-life of gentamicin of 2-5 h). The accuracy of the system has been
confirmed previously (Chambers et al., 1991) and during this study by measurement of
gentamicin concentrations from timed aliquots. Samples of broth were automatically
removed from the culture chamber each hour by a computer controlled peristaltic
pump, in order to correct volumes. Eluent was sampled hourly for 8 h, then at 10, 12,
13, 14, 18, 19, 20, 22, 24, 26, and 28 h, 100 /JL of the eluent was diluted as necessary
with sterile water, plated onto blood agar plates, and incubated at 37CC. The viable
count was determined after overnight growth (24-30 h). Dilution of the sample reduced
the concentration of gentamicin to undetectable levels.
Four regimens of gentamicin dosing were compared, each with a total gentimicin
AUC of 93-6 mg/L.h for a 24 h period (n = 4 for each regimen): (1) continuous
infusion at 3-9 mg/L; (2) peak 8-0 mg/L, trough 1-5 mg/L, dose interval 6 h; (3) peak
13-5 mg/L, trough 0-5 mg/L, dose interval 12 h; (4) peak 260 mg/L, trough 0-03 mg/L,
dose interval 24 h. Further comparisons were made between single doses of gentamicin
which achieve peak concentrations of 3-9, 8-0, 135, and 26-0 mg/L and between
continuous infusions at each of these concentrations (n = 4 in each group).
The number of colony-forming units at mfliirrmm kill and at 6, 12 and 24 h in the
different groups were compared using one-way analysis of variance with post-hoc

Downloaded from https://academic.oup.com/jac/article-abstract/29/4/427/825215

by guest
on 19 March 2018
 Comparison of grntanririn dose regimens 429

Figure 2. Viable counts of P. aeruginosa in samples from the in-vitro model of equal AUCs. Gentaxnkin
concentrations were 3-9 mg/L continuous (O) (regimen 1); peak 8-0 mg/L, trough 1-5 mg/L (A) (regnnen 2);
peak 13-5 mg/L, trough 05 mg/L (A) (regimen 3); peak 26-0 mg/L, trough 0-3 mg/L (D) (regimen 4);
control (no antibiotic) ( • ) . The times of further gentamidn administration are shown by arrow*. Vertical
bars show standard error.

analysis of the least significant difference. A P value of 005 was used to indicate a
significant difference.

Results
The killing of P. aeruginosa and regrowth patterns are shown in Figures 2 to 4.
In all experiments the initial bactericidal effect was greater after exposure to higher
peak concentrations of gentamidn. In the experiments with equal gentamidn AUCs
(Figure 2), there was a significant difference in the maximum decrease in viable counts
between each regimen with a successively greater maximum decrease from regimen 1
through to regimen 4. In the single dose experiments (Figure 3) there was a signifi-
cantly greater maximum decrease in viable counts as the peak concentration increased
from 3-9 mg/L to 26 g/L gentamidn. In the continuous infusion experiments
(Figure 4), the maximum decrease in viable counts was also significantly greater as the
concentration increased from 3-9 mg/L to 26 mg/L gentamidn.
The time taken for regrowth of the bacteria to numbers above the starting viable
counts in the experiments with equal AUCs (regimens 1 to 4) was prolonged in relation
to peak concentrations. After the numbers of bacteria had returned to starting values,
the rate of regrowth was similar for each regimen, and similar to that of the antibiotic-
free control regimen (Figure 2). The MIC of the final culture of P. aeruginosa NCTC
10662 was not determined, however in subsequent experiments with P. aeruginosa
AT 27853 the MIC of gentamidn for the resultant culture was unaltered (data not
shown).

Downloaded from https://academic.oup.com/jac/article-abstract/29/4/427/825215

by guest
on 19 March 2018
 430 E. J. Begg et aL

0 4 6 12 16 20 24 28
Tlmt ( h )
Figure 3. Viable counts of P. aerugtnosa in samples from the in-vitro model after tingle doses of
gentamicin. Peak gentamicin concentrations were 3-9 mg/L (O). 8-0 mg/L (A); 13-5 mg/L (A), 2fr0 mg/L.
(D)- Control (no antibiotic) (#). Vertical ban show standard error.

Subsequent doses of gentamicin had substantially less effect than the first dose
(Figure 2). In regimen 2, dosed every 6 h (peak 80, trough 1-5 mg/L), there was
minimal flattening of the slope of the regrowth curve after subsequent doses. In'
regimen 3, dosed every 12 h (peak 13-5, trough 0-5 mg/L) a secondary kill was evident,
after the dose at 12 h though not after the dose at 24 h. In regimen 4, dosed every 24 h
(peak 26-0, trough 0-03 mg/L) a secondary kill was evident after the second dose
though this was small compared with the first-dose effect. Regrowth also occurred in
every continuous infusion experiment despite the very large total doses administered
although some inhibitory effect was evident for the larger doses (Figure 4).
The duration of the post-antibiotic effect (PAE) was approximately 05 h for all
regimens, as denned by the time for one logarithm of growth to occur compared with
that of the control group at the time point at which the drug concentration fell below
the MIC (Figure 2). These times, (derived from t = lnQpeak]/MIC)/fc) were 35 h, 5-4 h
and 7-8 h for regimens 2, 3 and 4 with peaks of 80, 13-5 and 260 mg/L respectively.
This contrasted with a PAE of 1*7, 2-4 and 2-6 h measured by the standard washout
method (Vogelman & Craig, 1985), using gentamicin concentrations of 2-7, 6-7 and 10
times the MIC respectively.

Discussion
Kinetic in-vitro models enable drug concentrations to be controlled to mimic human
in-vivo concentrations. They most closely approximate conditions in neutropenic,
immunocompromised patients (Zinner et al., 1985) and have been shown to correlate
well with studies in neutropenic mice (Gerber & Keller-Segessenmann, 1985).

Downloaded from https://academic.oup.com/jac/article-abstract/29/4/427/825215

by guest
on 19 March 2018
 Comparison of gentwridn don regimen 431

Flgnre 4. Viable counts of P. aeruginosa in samples from the in-vitro model during continuous infusions of
gentamicin. Sustained concentrations of gentamicin were 3-9 mg/L (O), 8-0 mg/L (A); 13-5 mg/L (AX 260
mg/L ( • ) • Control (no antibiotic) ( # ) . Vertical bars show standard error.

P. aeruginosa has always been difficult to treat and there have been various attempts
to use kinetic in-vitro models to help choose an optimal treatment regimen
(Guggenbichler, Semenitz & Konig, 1985).
In this study, it was shown that larger aminoglycoside doses given less frequently
have substantially greater bactericidal activity and produce longer periods until
regrowth than conventional dose regimens. Ledergerber, Blaser & Luthy (1985)
reported similar findings in their in-vitro study of the effects of three dosing regimens of
gentamicin (peak concentrations of 87, 16, 32 mg/L, dose interval 8, 16 and 32 h
respectively) on P. aeruginosa. These regimens do not however have identical AUCs
over the entire dosing period, and the effects of continuous infusion were not observed.
Blaser et al. (1985a) also studied P. aeruginosa in an in-vitro dynamic model and
demonstrated that a single daily dose of netilmicin was as effective as a 50% greater
total dosage given in three divided doses every 8 h and achieved a greater initial kill.
The second and third doses had no demonstrable bactericidal effect. In another study
Blaser, Stone & Zinner (19856) using the same total daily dose of netilmicin showed
faster killing of Escherichia coli, Klebsiella pneumoniae, Staphylococcus aureus and
P. aeruginosa with single daily doses than with 8-hourly divided doses or a continuous
infusion. Regrowth was greater with P. aeruginosa than the other organisms at a given
concentration of aminoglycoside. If the aminoglycoside concentrations were adjusted
to give similar ratios of drug concentration to MIC, regrowth of S. aureus and E. coli
was remarkably similar to that of P. aeruginosa.
The fact that regrowth of P. aeruginosa occurred despite continuous infusions of
gentamicin sustaining concentrations of 8, 13-5 and 26 mg/L was remarkable. The rate
of regrowth was slower than that observed in the equal AUC experiments (Figures 2

Downloaded from https://academic.oup.com/jac/article-abstract/29/4/427/825215

by guest
on 19 March 2018
 432 E. J. Begg et aL

and 4) suggesting that high concentrations of aminoglycoside do exert some suppres-

sive effect on growth.
The data are consistent with the phenomenon of adaptive resistance, which describes
a loss of bactericidal activity following initial exposure to aminoglycosides (Daikos et
al., 1990). This effect is enhanced by continual exposure to aminoglycosides and
reverses some hours after the drug is removed from the growth medium. This
phenomenon has been confirmed in vivo (Daikos, Lolans & Jackson, 1991). The
mechanism appears to involve temporary down-regulation of the second energy-
dependent pathway of internalization of aminoglycoside into bacteria.
The minimal post-antibiotic effect demonstrated in our model may be the result of
the slow exponential decline in aminoglycoside concentrations rather than the
extremely rapid removal used in models designed specifically to look for this effect.
Kinetic models simulating human drug concentrations are, however, more relevant to
human dosing. Analysis of the data of Ledergerber et al. (1985) and Gerber &
Feller-Segessenmann (1985) also reveals minimal post-antibiotic effect. The logic of
extrapolating findings produced by rapid removal of antibiotic to the in-vivo situation
is open to question.
These studies support the proposition that large doses administered at longer dosing
intervals are more efficacious. A large first dose appears particularly important.
Toxicity might also be improved by this approach (Chan, 1989). Further clinical studies
are needed to determine optimum target peak and trough concentrations.

References
Blaser, J., Stone, B. B., Groner, M. C. & Zinner, S. H. (1985a). Impact of nctilmicin regimens on
the activities of ccflazklimc-nctilmicin combinations against Pseudomonas aeruginosa in an
in vitro pharmacolcinetic model. Antimicrobial Agents and Chemotherapy 28, 64-8.
Blaser, J., Stone, B. B. & Zinner, S. H. (19856). Efficacy of intermittent versus continuous
administration of netilmicin in a two-compartment in vitro model. Antimicrobial Agents and
Chemotherapy 27, 343-9.
Chambers, S. T., Peddie, B. A., Robson, R. A., Begg, E. J. & Boswell, D. R. (1991). Antibacterial
effects of lomefloxacin in-vitro. Journal of Antimicrobial Chemotherapy 27, 481-9.
Chan, G. L. C. (1989). Alternative dosing strategy for aminoglycosides: impact on efficacy,
nephrotoxicity, and ototoxknty. DICP, The Annals of Pharmacotherapy 23, 788-93.
Daikos, G. L., Jackson, G. G., Lolans, V. T. & Livermore, D. M. (1990). Adaptive resistance to
aminoglycoside antibiotics from first-exposure down regulation. Journal of Infectious
• Diseases 16, 414-20.
Daikos, G. L., Lolans, V. J. & Jackson, G. G. (1991). First exposure adaptive resistance to
aminoglycoside antibiotics in vivo with meaning for optimal clinical use. Antimicrobial
Agents and Chemotherapy 35, 117-23.
Gerber, A. U. & Feller-Segessenmann, C. (1985). In-vitro assessment of in-vitro killing patterns
of Pseudomonas aeruginosa. Journal of Antimicrobial Chemotherapy 15, Sup pi. A, 201-6.
Guggenbichler, J. P., Semenitz, E. & Konig, P. (1985). Kill kinetics and regrowth pattern of
bacteria exposed to antibiotic concentrations simulating those observed in vivo. Journal of
Antimicrobial Chemotherapy 15, Suppl. A, 139-46.
Ledergerber, B., Blaser, J. & Luthy, R. (1985). Computer-controlled in-vitro simulation of
multiple dosing regimens. Journal of Antimicrobial Chemotherapy 15, Suppl. A, 169-73.
Moore, R. D., Smith, C. R. & Lietman, P. S. (1984). The association of aminoglycoside plasma
levels with mortality in patients with Gram-negative bacteremia. Journal of Infectious
Diseases 149, 443-8.
Moore, R. D., Lietman, P. S. & Smith, C. R. (1987). Clinical response to aminoglycoside
therapy: importance of the ratio of peak concentration to minimal inhibitory concentration.
Journal of Infectious Diseases 155, 93-9.

Downloaded from https://academic.oup.com/jac/article-abstract/29/4/427/825215

by guest
on 19 March 2018
 Comparison of gentamJdn dose regimens 433

Noonc, P., Parsons, T. M. C , Pattison, J. R., Slack, R. C. B., Garfidd-Davies, D. & Hughes, K.
(1974). Experience in monitoring gentamicin therapy during treatment of serious gram-
negative sepsis. British Medical Journal i, 477-81.
Powell, S. H., Thompson, W. L., Lutbe, M. A., Stern, R. C , Grossniklaus, D. A., Bloxham, D.
D. et al. (1983). Once-dairy vs continuous aminoglycoside dosing: efficacy and toxicity in
animal and clinical studies of gentamirin, netilmicin, and tobramycin. Journal of Infectious
Diseases 147, 918-32.
Vogebnan, B. S. & Craig, W. A. (1985). Postantibiotic effects. Journal of Antimicrobial
Chemotherapy 15, Suppl. A, 37-46.
Zinner, S. H., Blaser, J., Stone, B. B. & GroneT, M. C. (1985). Use of an in-vitro kinetic model to
study antibiotic combinations. Journal of Antimicrobial Chemotherapy 15, Suppl. A, 221-6.

{Received 20 March 1991; revised version accepted 15 November 1991)

Downloaded from https://academic.oup.com/jac/article-abstract/29/4/427/825215

by guest
on 19 March 2018