Pharmacokinetics and Pharmacodynamics of Intravenous Levofloxacin at 750 Milligrams and Various Doses of Metronidazole in Healthy Adult Subjects

Eighteen healthy volunteers, between the ages of 18 and 45 years, participated in a three-way randomized crossover study after informed written consent was obtained. The protocol was approved by the University of Illinois at Chicago and Western Institutional Review Boards. All subjects were required to have a baseline medical history, physical examination, and laboratory evaluation (complete blood cell count with differential, serum chemistries, and urinalysis) within 2 weeks of study initiation. Metropolitan Life Insurance Tables were used to ensure that all subjects were within 15% of their ideal body weight (21).

Exclusion criteria included a history or clinical evidence of hepatic, renal, or gastrointestinal disease; presence of any clinically significant baseline laboratory abnormalities; ingestion of any fluoroquinolone antibiotic, metronidazole, antacids, theophylline, or other drugs known to interact with fluoroquinolone antibiotics or metronidazole during the 2 weeks before study drug administration; a history of allergy, hypersensitivity, or other serious adverse reactions to metronidazole, levofloxacin, ofloxacin, or any fluoroquinolone antibiotic; history of any seizure disorder; and the use of any tobacco products during the 12 months before study drug administration. Females were excluded if they had not undergone medical sterilization, hysterectomy, or tubal ligation.

The study was designed as a randomized, three-way crossover trial involving 18 healthy adult subjects. Ten subjects were admitted to the General Clinical Research Center at the University of Illinois at Chicago Medical Center on the morning of the study for three 72-h study periods. The remaining eight subjects were studied at the Pulmonary Associates Research Center in Phoenix, Ariz. Subjects were randomized to initially receive one of three combinations of intravenous levofloxacin and intravenous metronidazole: levofloxacin at 750 mg once daily plus metronidazole at 500 mg every 8 h, levofloxacin at 750 mg once daily plus metronidazole at 1,000 mg once daily, and levofloxacin at 750 mg once daily plus metronidazole at 1,500 mg once daily (Table 1). Each subject was then crossed over to receive the other two combinations. There was a minimum 4-day washout period prior to starting the next combination.

Levofloxacin was supplied as a premixed solution (Levaquin; 750-mg 5% dextrose injection; Ortho-McNeil Pharmaceutical, Inc., Raritan, N.J.). The metronidazole 500-mg dose was administered as premixed bags (metronidazole injection; USP RTU; Baxter Healthcare Corporation, Deerfield, Ill.). The metronidazole 1,000- and 1,500-mg doses were prepared by reconstituting vials of metronidazole powder (Flagyl powder for injection; SCS Pharmaceuticals, Chicago, Ill.) with sterile water for injection according to the manufacturer's recommendations and neutralizing them with sodium bicarbonate in Viaflex bags to a total volume of 220 and 230 ml, respectively.

During each of the three 72-h study periods, levofloxacin was administered as a 90-min intravenous infusion once daily for a total of three doses. After completion of the levofloxacin infusion, metronidazole was administered as a 60-min intravenous infusion. Metronidazole at 1,000 and 1,500 mg was administered once daily for a total of three intravenous doses. When metronidazole at 500 mg was administered every 8 h, subjects received nine intravenous doses in total.

Blood samples for pharmacokinetic and pharmacodynamic analysis were obtained through an indwelling peripheral catheter placed in the arm contralateral to that used for the infusion of levofloxacin and metronidazole. Blood and urine samples were collected for a 25.5-h period starting on the third day of each study period (48.0 to 73.5 h) for the measurement of levofloxacin, metronidazole, hydroxymetronidazole, and metronidazole acetic acid (for urine assay only). Blood samples for pharmacokinetic analysis were obtained at the following times for dosing regimens involving once-daily administration of levofloxacin and metronidazole: 48 (prior to administration of the third dose of levofloxacin), 49.5 (prior to the third dose of metronidazole and after the administration of the third dose of levofloxacin), 50, 50.5, 50.68, 50.86, 51.22, 52.5, 57.5, 60.5, 72.5, and 73.5 h. For the dosing regimen with once-daily levofloxacin and metronidazole administered every 8 h, blood samples for pharmacokinetic analysis were obtained at the following times: 48 (prior to administration of third dose of levofloxacin), 49.5 (prior to the seventh dose of metronidazole and after the administration of the third dose of levofloxacin), 50, 50.5, 50.68, 50.86, 51.22, 52.5, 57.5, 58.65, 60.5, 62, 65.5, 66.5, 70, 72.5, and 73.5 h. Bactericidal activity was determined in duplicate from blood samples obtained during the last dosing interval at 48.0 (levofloxacin trough), 49.5 (metronidazole trough, levofloxacin peak), 50.5 (metronidazole peak), and 52.5, 57.5, and 60.5 (midpoints) h by the microdilution method recommended by the NCCLS (22, 23). All plasma samples were collected in EDTA tubes and centrifuged immediately. Plasma was transferred and stored in four aliquots in polypropylene cryogenic vials at −70°C until analysis.

Serial urine samples were collected over 25.5 h for each study period between 48 and 73.5 h at the following intervals: 48 to 52, 52 to 56, 56 to 60, 60 to 64, and 64 to 73.5 h. The total volume of urine was recorded following each interval collection, and a 15-ml aliquot was removed for measurement of levofloxacin, metronidazole, hydroxymetronidazole, and metronidazole acetic acid. At the end of the 25.5-h collection interval, all urine collections were combined and a 15-ml aliquot was removed for measurement of creatinine in urine to calculate creatinine clearance (CL_CR) for each subject. All urine samples were stored in sterile polypropylene tubes and frozen at −70°C until analysis. All drug assays were performed at the Clinical Research Laboratory of the University of Illinois at Chicago College of Pharmacy.

Levofloxacin concentrations in plasma and urine were measured by a reversed-phase high-performance liquid chromatographic (HPLC) method based on the previously established procedure reported by Granneman and Varga (12). Modifications to the original assay procedure involved a change in the analytical column and the mobile-phase composition. These modifications were made to shorten the analysis time. The analytical reference standard for levofloxacin was obtained from R. W. Johnson, Spring House, Pa. The internal standard for the plasma assay, A-47084, a p-bromophenyl-substituted quinolone, was synthesized at Abbott Laboratories (Abbott Park, Ill.). Briefly, the HPLC system consisted of a Waters Alliance 2695 separations module (Waters Associates, Milford, Mass.) and a Spectroflow 980 programmable fluorescence detector (Applied Biosystems, Foster City, Calif.). The mobile phase was a mixture of acetonitrile-water (42:58, vol/vol) containing 69 mM phosphoric acid, 17 mM NaH₂PO₄, 24 mM sodium dodecyl sulfate, and 8.7 mM N-acetohydroxamic acid. The flow rate was 1.5 ml/min. Separation was achieved on a Symmetry C₁₈ column (5-μm particle size, 3.9 by 150 mm; Waters Associates) kept at 30°C. Fluorescence detection was performed at 280 nm (excitation) and 389 nm (emission). The retention times for levofloxacin and A-57084 (internal standard) were 3.7 and 9.8 min, respectively, with a total run time of 12 min (plasma assay) and 6 min (urine assay). Quantitation was done using peak areas for the urine assay, and peak area ratios (peak area of levofloxacin divided by peak area of internal standard) for the plasma assay (Millennium 32 Chromatography Manager; Waters Associates).

For the plasma assay, 200 μl of plasma was combined with 200 μl of internal standard (4 μg of A-57084/ml) in displacing reagent (acetonitrile-water containing 75 mM Na₂HPO₄, adjusted to pH 7.4, and 17 mM sodium dodecyl sulfate, 30:70, vol/vol). This mixture was transferred to a Centrifree YM-30 micropartition device (Amicon Bioseparations, Millipore Corporation, Bedford, Mass.) and centrifuged using a fixed-angle rotor for 25 min. The ultrafiltrate was transferred to a limited-volume HPLC vial, and a 10-μl portion was analyzed.

The standard curve for levofloxacin in plasma was linear (r² ≥ 0.99) over the concentration range of 0.092 to18.0 μg/ml. The intraday coefficients of variation for replicate quality control (QC) samples (n = 5) within the curve range varied from 1.2 to 2.5%. The interday coefficients of variation were from 1.8 to 4.3%. The lower limit of detection for levofloxacin in plasma was 0.092 μg/ml.

For the urine assay, study samples were diluted 100-fold with deionized water. An aliquot of the diluted sample was transferred to an autosampler vial, and a 10-μl portion was analyzed.

The standard curves for levofloxacin in urine were linear (r² ≥ 0.99) over the concentration range of 0.101 to 20.3 μg/ml. The intraday coefficients of variation for replicate QC samples (n = 7) within the curve range varied from 1.1 to 2.4%. The interday coefficients of variation varied from 1.2 to 2.6%. The lower limit of detection for levofloxacin in urine was 0.101 μg/ml.

Metronidazole and its metabolites were measured by an HPLC method based on a previously established method (14). The HPLC method was further modified by changing the column from a μBondapak C₁₈ to a Symmetry C₁₈ column and in turn altering the mobile phase. These changes were made to improve the chromatography. The analytical reference standard for metronidazole was obtained from United States Pharmacopeia (Rockville, Md.). Standards for hydroxymetronidazole and metronidazole acetic acid were synthesized at SynChem Inc. (Des Plaines, Ill.). Tinidazole, the internal standard, was purchased from Sigma Chemical Company (St. Louis, Mo.). Briefly, the HPLC system consisted of a Waters Alliance 2695 separation module and 2487 dual absorbance detector (Waters Associates). The mobile phase was a mixture of acetonitrile and 0.02 M sodium acetate, pH 5.0 (4:96, vol/vol). The flow rate was 1.4 ml/min. Separation was achieved on a Symmetry C₁₈ column (5-μm particle size, 3.9 by 150 mm) kept at 30°C. UV detection was performed at 324 nm. The retention times were approximately 1.8, 3.6, 6.6, and 17.3 min for metronidazole acetic acid, hydroxymetronidazole, metronidazole, and internal standard, respectively, with a total run time of 19 min. Quantitation was done using peak area ratios, peak area of metronidazole, or metabolites divided by peak area of internal standard (Millennium 32 Chromatography Manager; Waters Associates). Metronidazole, hydroxymetronidazole, and metronidazole acetic acid concentrations were measured in plasma, while only metronidazole and its hydroxy metabolite were quantitated in urine.

The plasma sample preparation was adapted from the method used by Granneman and Varga (12). For this assay, 200 μl of plasma was combined with 200 μl of internal standard (16 μg of tinidazole/ml) in displacing reagent (acetonitrile-water containing 75 mM Na₂HPO₄, adjusted to pH 7.4, and 17 mM sodium dodecyl sulfate, 30:70, vol/vol). This mixture was transferred to a Centrifree YM-30 micropartition device and centrifuged using a fixed-angle rotor for 25 min. The ultrafiltrate was transferred to a limited-volume HPLC vial, and a 50-μl portion was analyzed.

For the first five subjects, the standard curves for metronidazole and hydroxymetronidazole in plasma were linear (r² ≥ 0.99) in the range of concentrations from 1.99 to 52.8 and 2.00 to 53.4 μg/ml, respectively. The interday coefficients of variation were 1.65 to 1.91% for metronidazole and 1.41 to 2.39% for hydroxymetronidazole. The lower limits of detection were 1.99 μg/ml for metronidazole and 2.00 μg/ml for hydroxymetronidazole.

For subjects 6 through 18, two lower calibrators were added on to the original curve and the assay was revalidated. The standard curves were linear (r² ≥ 0.99) in the range of concentrations of 0.498 to 52.8 μg/ml for metronidazole in plasma and 0.500 to 53.4 μg/ml for hydroxymetronidazole in plasma. The interday coefficients of variation ranged from 1.12 to 4.62 and 1.37 to 4.61% for metronidazole and hydroxymetronidazole, respectively. The lower limits of detection were 0.513 μg/ml for metronidazole and 0.500 μg/ml for the hydroxy metabolite.

The urine sample preparation followed the procedure described by Jensen and Gugler (14). For this assay, 100 μl of urine was combined with 400 μl of 0.2 M sodium acetate, pH 4.5, and 50 μl of 1,200-U/ml β-glucuronidase (type H-5 from Helix pomatia; Sigma Chemical Company). The sample was incubated at 37°C for 16 h. Following incubation, 50 μl of internal standard (1.85 mg of tinidazole/ml) was added to the sample. The mixed sample was transferred to an autosampler vial, and a 10-μl portion was analyzed. For this assay, calibrators were prepared in acetate buffer and 50 μl of acetate buffer was substituted for 50 μl of β-glucuronidase. In addition, the calibrators were not incubated.

For urine concentrations of metronidazole, hydroxymetronidazole, and metronidazole acetic acid, the standard curves were linear (r² ≥ 0.99) over the range of concentrations of 20.9 to 502, 21.7 to 519, and 10.9 to 522 μg/ml, respectively. The intraday coefficients of variation for replicate QC samples (n = 5) within these concentration ranges varied from 1.5 to 3.6% for both the parent and metabolites. The interday coefficients of variation ranged from 2.0 to 2.5% for metronidazole, 2.3 to 2.4% for hydroxymetronidazole, and 2.2 to 3.5% for metronidazole acetic acid. The lower limits of detection were 20.9 μg/ml for metronidazole, 21.4 μg/ml for hydroxymetronidazole, and 10.9 μg/ml for metronidazole acetic acid.

The creatinine assay was performed using a commercially available kit (Direct Creatinine Reagent Set, catalog no. 2600; Eagle Diagnostics, De Soto, Tex.) and read on a Spectronic 21 spectrophotometer (Milton Roy). Urine samples were diluted 10-fold with deionized water in accordance with the protocol and assayed at 510 nm. According to assay literature, the procedure is linear in the range of 0 to 20 mg/dl. A calibrator (6.0 mg/dl, provided with the kit) and reagent blank were determined with each set of unknowns tested. The ratio of unknown absorbance to calibrator absorbance was used to determine unknown concentration. Normal and elevated QC samples (Chem-trol normal, 4.0 mg/dl, catalog no. 8100, and Chem-trol elevated, 9.5 mg/dl, catalog no. 8200; Eagle Diagnostics) were analyzed with each run.

The amount of creatinine recovered from the 25.5-h urine collection was determined and used to calculate a measured CL_CR for each subject during each study period. The expected daily creatinine production for male subjects was determined to verify an accurate urine collection.

Noncompartmental analysis was used to estimate pharmacokinetic parameters for levofloxacin, metronidazole, and hydroxymetronidazole by using the microcomputer program WinNonlin (version 3.1; Pharsight Corporation, Mountain View, Calif.). The maximum (C_max) and minimum (C_min) concentrations were determined by visual inspection of the observed concentration-versus-time curve for each subject. Within a given dosing interval (τ), the area under the plasma concentration-time curve (AUC_0-τ) and area under the first-moment curves (AUMC_0-τ) were calculated using the log-linear trapezoidal rule. Systemic exposure was reported as AUC_0-24 for doses of levofloxacin at 750 mg, metronidazole at 1,000 mg, and metronidazole at 1,500 mg. When metronidazole was administered every 8 h, AUC_0-τ or AUC_0-8 was multiplied by 3 and reported as AUC_0-24. Total body clearance (CL_ss) was calculated from the equation CL_ss = dose/AUC_0-τ. Log-linear regression of the terminal slope was used to estimate the elimination rate constant (β). The elimination half-life (t_1/2) was calculated by dividing β into the natural logarithm of two. The volume of distribution at steady state (V_ss) was calculated from the equation V_ss = (mean residence time [MRT] extrapolated to infinity) × (CL_ss), when MRT = ([AUMC_0-τ + τ(AUC_0-∞ − AUC_0-τ)]/AUC_0-τ) − (T_inf/2). AUC_0-∞ is the AUC up to the last measurable concentration extrapolated to infinity by adding on the area determined from the last measurable concentration in plasma divided by β. T_inf/2 is the average time that a molecule remains in the infusion set during a constant infusion (T_inf), or the mean input time (MIT).

The area under the plasma concentration-versus-time curve from time zero to the last measurable concentration in plasma (AUC_last) was determined using the log-linear trapezoidal rule. The renal clearance (CL_R) of levofloxacin was calculated from the cumulative amount of drug excreted in the urine (Ae) during the 25.5-h collection period divided by the AUC from time zero to the last measurable plasma levofloxacin concentration (AUC_last or AUC_0-25.5). Since the metronidazole infusions were started after completion of the levofloxacin infusion, the plasma AUC_last is reflective of only 24 h (49.5 to 73.5 h) of the 25.5-h urine collection period (48.0 to 73.5 h). The trapezoidal rule was utilized to calculate the 1.5 h (48.0 to 49.5 h) of area prior to the start of the metronidazole infusion. For the once-daily regimens of metronidazole, this area was added to the AUC_last (approximately 24 h) of metronidazole and hydroxymetronidazole in order to reflect the plasma concentration-time exposure over 25.5 h (AUC_0-25.5). When metronidazole was administered every 8 h, the AUC_last (approximately 8 h) was multiplied by 3 and then added to the 1.5 h of area prior to the seventh dose of metronidazole to determine the AUC_0-25.5. The CL_R values of metronidazole and hydroxymetronidazole were calculated by dividing the Ae by the corresponding plasma AUC_0-25.5.

Body surface area was determined for all subjects using the DuBois equation (10). Both CL and CL_R values were standardized for each subject to a body surface area of 1.73 m².

The MIC and minimal bactericidal concentrations (MBC) of levofloxacin and metronidazole were determined for clinical isolates of E. coli, B. fragilis (n = 2), B. thetaiotaomicron, and P. asaccharolyticus by the microdilution methods recommended by the NCCLS (22, 23). The E. coli isolate was obtained from the clinical microbiology laboratory at the University of Illinois at Chicago Medical Center. The anaerobic organisms were obtained from David Hecht (Loyola University Hospital, Maywood, Ill.). E. coli ATCC 25922 and B. fragilis ATCC 25285 were included for validation of MIC results. All organisms underwent three passages on nonselective medium prior to testing. Direct suspensions of the organisms were made from 24- and 48-h growth of E. coli and the anaerobes, respectively. E. coli and the Bacteroides species were prepared in sterile normal saline to match a 0.5 McFarland turbidity standard at a spectrophotometric wavelength of 625 nm, with an absorbance of 0.08 to 0.1. Suspensions of P. asaccharolyticus were prepared at a slightly higher optical density (0.2). The final inoculum of E. coli for all assays was approximately 5 × 10⁵ CFU/ml. The final inoculum for the anaerobic organisms was approximately 10⁶ CFU/ml. Plate counts were performed in every assay to determine the exact inoculum density.

Stock solutions of the antibiotics were prepared according to NCCLS guidelines (22, 23) and stored at −80°C until used. Serial twofold dilutions of the antibiotics were prepared in cation-adjusted (Ca²⁺, 25 μg/ml; Mg²⁺, 12.5 μg/ml) Mueller-Hinton broth (Difco, Detroit, Mich.) for aerobic MICs and in brucella broth (Difco) supplemented with vitamin K and hemin (BBL, Becton Dickinson and Company, Sparks, Md.) and lysed horse blood (Hema Resources and Supply, Inc., Aurora, Oreg.) for anaerobic MICs. Final antibiotic concentrations tested were 0.125 to 64 μg/ml for metronidazole and 0.015 to 8 μg/ml for levofloxacin.

The MIC of levofloxacin for E. coli was determined after 24 h of incubation at 35°C. The MICs of levofloxacin and metronidazole for the anaerobic isolates were read after 48 h of incubation at 35°C in an anaerobic chamber (Bactron anaerobic chamber; Sheldon Manufacturing, Inc., Cornelius, Oreg.). The MICs were read as the lowest antimicrobial concentration showing no visible turbidity in the microtiter plate wells. For MBC determinations, a 25-μl sample was removed from each clear well and inoculated onto a blood agar plate (for E. coli) or a CDC anaerobic blood agar plate (for anaerobes). After the sample dried, the plates were streaked in three planes and incubated at 35°C in an appropriate environment. Colonies of E. coli were counted after 24 h of incubation, with anaerobic viable counts determined after 48 h of incubation. The MBC endpoint was defined as the plate showing >99.9% killing of the inoculum.

Inhibitory (SIT) and bactericidal (SBT) titers ranging from 1:2 to 1:1,024 were determined using serial twofold dilutions of human plasma and appropriate broth media. Bacterial suspensions were prepared as described previously. SIT and SBT endpoints were determined as described previously for MIC and MBC assays. All procedures were performed in duplicate. The 24-h areas under the inverse serum inhibitory curve (AUIC_0-24) and bactericidal titer curve (AUBC_0-24) were determined by the trapezoidal rule with the microcomputer program WinNonlin (version 3.1).

The pharmacokinetics and pharmacodynamic parameters were summarized by presenting the mean and standard deviation (SD) for each treatment group. Analysis of variance methods were used to assess significant differences among the three metronidazole groups by using SAS-PROC GLM. The analysis compares treatments for each of the parameters of interest in the log scale. The parameters for each subject were log-transformed prior to assessing differences between treatments. An analysis of variance model on the log-transformed data with effects for sequence group, treatment, period, and subject-within-sequence was used for this crossover design. Pairwise treatment comparisons were carried out using the Student-Newman-Keuls test. Statistical significance was determined at the P ≤ 0.05 level. Tests for the normality and equality of variances were performed with Shapiro-Wilks and Levene tests, respectively. As a result, all parameters of interest were logarithmically transformed to effectively normalize the data. All data analyses were performed using the statistical software package PC SAS (version 8.2; SAS Institute, Cary, N.C.).

Mean pharmacokinetic parameters of levofloxacin and metronidazole for each dosing regimen are presented in Table 2. The mean pharmacokinetic parameters of the hydroxy metabolite of metronidazole are listed in Table 3. The pharmacokinetic parameters for levofloxacin were not significantly different between study periods. No sequence effects for the pharmacokinetic parameters were observed. No carryover effects were noted, except for the following instances: metronidazole t_1/2 (1,000 mg once daily), metronidazole CL_R (ml/min/1.73 m², 1,000 mg once daily), metronidazole CL_R (ml/min, 500 mg every 8 h and 1,000 mg once daily), and AUIC_0-24 against B. thetaiotaomicron. The carryover effect did not affect the overall treatment P value.

When metronidazole was administered as a 1,500-mg dose once daily, the C_max (37.7 ± 10 mg/liter) was significantly higher than those of the other two regimens (P < 0.05). The C_max values observed with the 1,000-mg once-daily and 500-mg every-8-h regimens were not significantly different from each other (24.8 ± 6.8 versus 22.2 ± 5.0 mg/liter, respectively). The C_min observed when metronidazole was given as a 500-mg dose every 8 h was 10. 4 ± 1.9 mg/liter and was significantly higher than those of the other two regimens (P < 0.05). The C_min for metronidazole at 1,500 mg once daily (4.8 ± 1.4 mg/liter) was significantly higher (P < 0.05) than that observed when metronidazole was given as a 1,000-mg once-daily dose (3.4 ± 1.3 mg/liter). Comparison of the AUC_0-24 values for metronidazole at 1,500 mg administered once daily (338 ± 105 mg · h/liter) and metronidazole at 500 mg administered every 8 h (356 ± 68 mg · h/liter) revealed no significant difference (P > 0.05). The AUC_0-24 for the metronidazole 1,000-mg once-daily regimen (227 ± 57 mg · h/liter) was significantly lower (P < 0.05) than those of the other two dosing regimens. The half-life (t_1/2) observed with the 500-mg every-8-h regimen was 8.0 ± 1.3 h and was significantly less than the t_1/2 observed when metronidazole was administered once daily (P < 0.05). The steady-state volume of distribution (V_ss) was significantly higher with the 1,000- and 1,500-mg once-daily regimens (59.0 ± 10 and 62.5 ± 11 liters, respectively) compared to the metronidazole 500-mg every-8-h regimen (49.2 ± 12 liters). Systemic clearance (CL) and renal clearance (CL_R) were not significantly different among the three metronidazole dosing regimens. No difference was observed in the cumulative amount of metronidazole, hydroxymetronidazole, or metronidazole acetic acid that was excreted in the urine (Ae) between the 1,500-mg once-daily and 500-mg every-8-h regimens (Table 4). However, the cumulative amount of metronidazole, hydroxymetronidazole, and metronidazole acetic acid excreted in the urine during the 1,000-mg once-daily regimen was significantly lower than those observed with the other two regimens. Over 25.5 h, the mean percentages of a dose that were recovered in the urine as metronidazole, hydroxymetronidazole, and metronidazole acetic acid for each dosing regimen were as follows: 500 mg every 8 h (19, 31.2, and 18.2%, respectively), 1,000 mg once daily (16.8, 25.8, and 15%, respectively), and 1,500 mg once daily (17.3, 28.3, and 16%, respectively).

Levofloxacin and metronidazole were well tolerated in the 18 subjects completing all phases of the study, and no serious drug-related adverse effects were noted. Fifteen subjects reported one or more mild probably or possibly drug-related adverse effects. These effects included metallic or unpleasant taste (n = 12), irritation at the site of infusion (n = 8), nausea (n = 2), diarrhea (n = 2), and headache (n = 1) in 14 subjects during the levofloxacin 750-mg and metronidazole 1,500-mg once-daily regimen. Eight subjects during the metronidazole 1,000-mg and levofloxacin 750-mg once-daily regimen reported an unpleasant or metallic taste (n = 6), irritation at the site of infusion (n = 2), and headache (n = 1). For the metronidazole 500-mg every-8-h and levofloxacin 750-mg once-daily regimens, adverse effects in 13 subjects included metallic or unpleasant taste (n = 10), infusion site irritation (n = 5), headache (n = 3), and diarrhea (n = 3). During the prestudy evaluation, clinical laboratory tests were within the normal range for all subjects. No clinically significant changes in laboratory values were observed during the poststudy evaluation.

The MICs and MBCs of levofloxacin for the organisms tested were as follows: B. fragilis 1 (4.0 and 8.0 μg/ml), B. fragilis 2 (4.0 and 8.0 μg/ml), B. fragilis ATCC 25285 (2.0 and 8.0 μg/ml), B. thetaiotaomicron (>8.0 and >8.0 μg/ml), P. asaccharolyticus (8.0 and 8.0 μg/ml), E. coli (0.125 and 0.125 μg/ml), and E. coli ATCC 25922 (0.06 and 0.06 μg/ml, respectively). For metronidazole the MICs and MBCs for the organisms tested were as follows: B. fragilis 1 (1.0 and 1.0 μg/ml), B. fragilis 2 (1.0 and 2.0 μg/ml), B. fragilis ATCC 25285 (0.5 and 1.0 μg/ml), B. thetaiotaomicron (2.0 and 2.0 μg/ml), and P. asaccharolyticus (1.0 and 1.0 μg/ml, respectively).

Median reciprocal inhibitory titers and measured inhibitory activities (AUIC_0-24) for the three dosing regimens are reported in Table 5. Median reciprocal bactericidal titers (data not shown) and measured bactericidal activities (AUBC_0-24) were very similar to the values reported for inhibitory titers and AUIC_0-24. The respective mean (± SD) AUBC_0-24 values for the dosing regimens of levofloxacin at 750 mg once daily plus 500 mg every 8 h, 1,000 mg once daily, or 1,500 mg once daily were as follows: B. fragilis 1, 379 ± 95, 293 ± 83, and 367 ± 100; B. fragilis 2, 260 ± 76, 193 ± 55, and 263 ± 76; B. thetaiotaomicron, 204 ± 78, 169 ± 53, and 220 ± 77; P. asaccharolyticus, 617 ± 240, 379 ± 138, and 606 ± 202; and E. coli, 1,953 ± 524, 1,974 ± 544, and 1,961 ± 605, respectively.

Against E. coli, the AUIC_0-24 and AUBC_0-24 were not significantly different between the regimens. For the anaerobic pathogens, there was no significant difference in the AUIC_0-24 and AUBC_0-24 between the regimen of levofloxacin at 750 mg once daily plus metronidazole at 500 mg every 8 h and the regimen of levofloxacin at 750 mg once daily plus metronidazole at 1,500 mg once daily, with the exception of an AUIC_0-24 value against B. thetaiotaomicron. The AUIC_0-24 for the regimen of levofloxacin at 750 mg and metronidazole at 1,500 mg once daily against B. thetaiotamicron was significantly higher (P ≤ 0.05) than those of the other two dosing regimens.

The values of AUIC_0-24 and AUBC_0-24 for the regimen of levofloxacin at 750 mg plus metronidazole at 1,000 mg once daily were significantly (P < 0.05) less than those of the other two regimens against the anaerobic pathogens, with the exception that there was no significant difference found for the AUIC_0-24 against B. thetaiotamicron between metronidazole 1,000-mg once-daily and metronidazole 500-mg every-8-h regimens when administered with levofloxacin.