TEXT

Top Abstract Text References

Levofloxacin is the active L-isomer of the racemate ofloxacin, a fluoroquinolone with a broad spectrum of activity (5, 14). We assessed the pharmacokinetics of levofloxacin and the serum bactericidal activities (SBAs) against five different species of members of the family Enterobacteriaceae. Twelve healthy female volunteers (mean age, 33.8 ± 6.1 years; mean body weight, 64.1 ± 10.2 kg) were included after they provided written informed consent. Exclusion criteria were hypersensitivity to quinolones, pregnancy, drug or alcohol abuse, or the use of a concomitant medication except for contraceptives. The study was approved by the ethical committee of the Physicians' Association of Hessen, Frankfurt, Germany.

Levofloxacin (lot HR355/1014, batch no. 20; Hoechst, Frankfurt, Germany) was administered orally as an open single dose of 500 mg (in tablet form) on an empty stomach after 10 h of fasting. Blood samples were collected just before medication and at 0.5, 1, 2, 4, 6, 8, 12, and 24 h after medication. Serum levofloxacin concentrations were determined enantioselectively by high-performance liquid chromatography by a previously described method (12). The assay was linear over the concentration range studied, and the lower limit of quantitation was 20 ng/ml. The inter- and intraday coefficients of variation were both <3%.

We analyzed 10 clinically relevant strains of five different species of the family Enterobacteriaceae isolated by the Department of Clinical Microbiology of the City Hospital Zehlendorf/Heckeshorn and of the Departments of Microbiology of the University Hospitals Benjamin-Franklin and Charité-Virchow, Berlin, Germany. The species of all isolates were determined with the API 20E system (API BioMérieux, Nürtingen, Germany). The minimal bactericidal concentrations (MBCs) of levofloxacin were determined in triplicate by the standardized microdilution method described in the guidelines of the National Committee for Clinical Laboratory Standards (NCCLS) (16). All samples were serially diluted in a 96-well plate (Falcon; Becton Dickinson, Lincoln Park, N.J.) from a titer of 1:4 to a titer of 1:512 with an automatic microdilutor (Titertek; Denley, Billinghurst, United Kingdom). The diluent was Mueller-Hinton broth without Mg²⁺ and Ca²⁺ (Difco Laboratories, Detroit, Mich.). Every well was inoculated with the final inoculum of bacteria (5 × 10⁵ CFU/ml). The trays were incubated at 35°C for 24 h, and the wells were assessed for visible turbidity. One microliter from each well was spotted with a multipoint inoculator (Microtiter; Denley) on Mueller-Hinton agar plates, which were then incubated for 24 h at 35°C. The concentration of the dilution titer at which no bacterial growth occurred was considered the MBC. SBAs were determined in triplicate by the standardized microdilution method described in the NCCLS guidelines (15), with the dilution procedure for the sera being similar to that used for the MBC method. As the diluent, a mixture of equal parts of Mueller-Hinton broth without Mg²⁺ and Ca²⁺ and heat-inactivated (56°C, 30 min) pooled human serum was used. Pseudomonas aeruginosa ATCC 27853 (MIC, 1 to 4 mg/liter) and Escherichia coli ATCC 25922 (MIC, 0.016 to 0.06 mg/liter) were used as quality control strains.

Pharmacokinetics were analyzed compartmentally and noncompartmentally. All pharmacokinetic parameters were normalized to a body weight of 70 kg before averaging; the clearance values were normalized to a body surface of 1.73 m². The maximum concentration (C_max), the time to C_max (T_max), and the elimination half-life (t_1/2) were calculated by assuming an open one-compartment model for extravascular application. The decision was based on the Schwarz criterion (22). An iterative least-squares method was used to fit the regression curve to the experimentally obtained values. Nonlinear regression analysis was performed to minimize the following objective function:
where C_i is the concentration measured at time t_i (i = 1, ... n), and C(t) is the concentration measured at time t.

The area under the serum concentration-time curve (AUC) was calculated noncompartmentally by the log trapezoidal summation method (24). The apparent terminal half-life was calculated from the adjustment of a single exponential function (Ce^zt) to the terminal phase of the concentration-time profiles, where C is a constant and z is the apparent terminal rate constant. The adjustments were done by using the method of least squares. The half-lives were then calculated as t_1/2;z = ln(2)/z, where t_1/2;z is the terminal half-life. The relative total volume of distribution at steady state (V_ss/) and the clearance (CL/) were calculated noncompartmentally, where  is the bioavailability. SBA data are indicated as geometric mean titers and as the mean ± standard deviation (SD) of the log₂ of all reciprocal serum bactericidal titers. Spearman's rank order correlation coefficient (r_s,t) corrected for ties (18) was used to analyze the correlation between C(t)/MIC and the corresponding SBA.

Pharmacokinetic data are presented in Table 1. There were only mild adverse effects in four volunteers (mild peripheral edema and mild headache); no abnormal physical or laboratory findings were observed. The MBC of levofloxacin was 0.008 mg/liter for all isolates of E. coli. For Klebsiella pneumoniae, the MBCs were 0.016 mg/liter (5 strains), 0.25 mg/liter (four strains), and 1 mg/liter (one strain). For Enterobacter cloacae, the MBCs were 0.008 mg/liter (five strains), 0.5 mg/liter (three strains), 1 mg/liter (one strain), and 4 mg/liter (one strain). For Serratia marcescens, the MBCs were 0.063 mg/liter for all strains. For Citrobacter freundii, the MBCs were 0.008 mg/liter (eight strains), 0.016 mg/liter (one strain), and 0.063 mg/liter (one strain). The MBCs for reference strains P. aeruginosa ATCC 27853(MBC, 2 mg/liter) and E. coli ATCC 25922 (MBC, 0.016 mg/liter) were within the acceptable quality control ranges of NCCLS (16). The SBA titers are indicated in detail in Table 2.

There was a strong positive correlation between the concentration in serum/MIC ratio and the corresponding SBA at l, 2, 12, and 24 h after drug administration (r_s,t = 0.846 for 2300 datum pairs; P < 0.0001). Similarly, the C_max/MIC ratio was strongly correlated with the SBA at 1 h after drug administration (r_{s, t} = 0.808 for 600 datum pairs; P < 0.001). The AUC from time zero to 24 h (AUC_0-24)/MIC ratio (indicated as inverse serum inhibitory titer [SIT] integrated over time) had a strong correlation with the area under the reciprocal of the SBA curve (r_s,t = 0.882 for 540 datum pairs; P < 0.0001).

Our pharmacokinetic data correspond to published data that indicate a C_max of 5.2 to 5.9 mg/liter within 1 to 2 h after administration of 500 mg of levofloxacin and a t_1/2;z of 6 to 8 h (2, 3, 5, 7, 11, 14). The bactericidal activities of levofloxacin are within the range of those of other quinolones, with MICs at which 90% of isolates are inhibited of 1 mg/liter for most members of the family Enterobacteriaceae (1, 8, 9, 17).

The SBA test has been used for many years to monitor antimicrobial activity in vivo (20, 25). SBA is supposed to occur at concentrations of 4× the MIC or greater, although patients with reduced immunity may require higher SBA titers of 8× to 16× the MIC (10, 17, 23). Our data confirm that SBA against members of the family Enterobacteriaceae occurs at a concentration in serum/MIC ratio of 4.

The antimicrobial effects of fluoroquinolones are concentration dependent (4, 13). An AUC_0-24/MIC ratio of >125 SIT¹ · h is accepted to be the significant breakpoint for probabilities of clinical and microbiological cures (13). However, there is evidence that the C_max/MIC ratio may be better correlated with the antimicrobial effect than the AUC_0-24/MIC ratio, with the breakpoint being a C_max/MIC ratio of 13:1 to 20:1 (6, 19, 21). Our data show that both the C_max/MIC ratio and the AUC_0-24/MIC ratio are significantly correlated with the SBA of levofloxacin.