Antibacterial Properties of Gemifloxacin and Trovafloxacin in Urine Ex Vivo: Phase I Study

Abstract

Urine bactericidal titers (UBTs) against Escherichia coli ATCC 25922 and Staphylococcus saprophyticus ATCC 1970 were determined after the administration of single oral doses of gemifloxacin at 320 mg and trovafloxacin at 200 mg to healthy volunteers. Gemifloxacin presented significantly lower experimental versus mathematically predicted UBTs over 72 h, due to the effect of urine on the susceptibility of the E. coli strain. Experimental UBTs were significantly higher for gemifloxacin than trovafloxacin against both strains over 72 h.

“Ex vivo” activity in urine is important since the measurement of urine antibacterial activity correlates directly with the outcome of infection (4) more than the determination of MICs and the levels of antimicrobials in urine (2), probably because MICs are higher when tested in human urine than in conventional media (6).

Urine bactericidal titers (UBTs) and the area under the urine bactericidal curve (AUBC) of gemifloxacin versus trovafloxacin were evaluated against Escherichia coli and Staphylococcus saprophyticus strains.

Twelve healthy male volunteers (mean age, 26.2 ± 3.5 years; height, 175.2 ± 4.7 cm; weight, 73.2 ± 4.2 kg) received single oral doses of 320 mg of gemifloxacin (SmithKline Beecham Pharmaceuticals, Harlow, United Kingdom) and 200 mg of trovafloxacin (Trovan; Pfizer Inc., New York, N.Y.), separated by a 14-day washout period. The protocol was approved by the Research Ethics Committee of Hospital La Paz, Madrid, Spain. Written informed consent was obtained from all subjects.

Urine samples were collected prior to drug administration (0 h) and at intervals of 0 to 2, 2 to 4, 4 to 8, 8 to 12, 12 to 16, 16 to 24, 24 to 36, 36 to 48, 48 to 60, and 60 to 72 h after dosing for determination of urine concentrations of gemifloxacin and trovafloxacin by high-pressure liquid chromatography (HPLC) and bioassay and for determination of urine bactericidal titers (UBTs).

In vitro susceptibility testing was performed using Laboratory reference standards of gemifloxacin and trovafloxacin (SmithKline Beecham Pharmaceuticals, Tonbridge, United Kingdom), five times each for E. coli ATCC 25922 and Staphylococcus saprophyticus ATCC 1970 by standardized methods (5). In addition, the same methodology was followed but using a pool of the volunteer's urine as test media to study the effect of urine on the antibacterial activity of the study drugs.

Gemifloxacin and trovafloxacin concentrations were determined by bioassay in plates with nutrient agar (BBL; Becton Dickinson, Cockeysville, Md.) seeded with Bacillus subtilis NCTC 6633 spore suspension (Difco Laboratories, Detroit, Mich.). A total of 90 μl of each sample was deposited into 7-mm-diameter wells in the inoculated plates and incubated at 30°C for 18 h. Gemifloxacin and trovafloxacin standards of from 32 to 0.125 μg/ml were prepared in pooled urine obtained from volunteers before administration of the drugs. The lower limit of detection was 0.125 μg/ml.

For HPLC determination, frozen samples were thawed at ambient temperature, agitated at 37°C using ordinary mixing for 30 min, and centrifuged before sampling. Gemifloxacin and trovafloxacin were extracted from human urine by using a dilution method. Gemifloxacin and trovafloxacin samples were analyzed by HPLC-MS/MS operated in the positive ion mode, using a Turbo IonSpray interface. Upon using a 50-μl aliquot, the lower limits of quantification for gemifloxacin and trovafloxacin in urine were determined to be 0.01 and 0.0250 μg/ml, respectively. Three quality control samples from each of three concentrations within the calibration range were assayed with each batch of experimental samples. The results of the quality control samples were used to assess the day-to-day performance of the assay.

UBTs were determined against reference strains E. coli ATCC 25922 and S. saprophyticus ATCC 1970 in the logarithmic and stationary phases of growth, respectively, by the microdilution technique (1). Urine samples from each volunteer were diluted in urine obtained from each volunteer before drug administration. The final volume of each well was 100 μl. The final inoculum was ca 10⁶ CFU/ml (5 × 10⁴ CFU/well). The inoculated plates were incubated at 35°C for 18 h and were subcultured in antibiotic-free blood supplemented Mueller-Hinton agar (bioMérieux, Mercy l'Etoile, France), incubated at 37°C for 18 h. The bactericidal endpoints were defined as the highest dilution of urine killing 99.9% of the original inocula.

Correlations between concentrations determined by HPLC and by bioassay were investigated by linear regression.

Areas under the concentration-time curve (AUCs) were calculated from a plot of concentrations of gemifloxacin or trovafloxacin determined in urine by bioassay versus the mid-time point of each collection period. The trapezoidal rule was used for the calculation. To measure the effect of urine in the ex vivo bactericidal activity, the AUC_{0–72 h}/MBC value was calculated for each subject with both drugs and strains using both the MBC value determined with urine and the MBC value determined in broth.

Predicted bactericidal titers were calculated using HPLC and bioassay concentrations by the method described by Drusano et al. (3) and displayed as 1/2ⁿ, where n is the number of halvings (dilutions) for which the resultant antimicrobial concentration remained above the MBC. Experimental and predicted AUBCs were calculated from a plot of experimental and predicted urine bactericidal titers versus the mid-time point of each collection period. Since the highest measurable bactericidal titer was ≥1,024, AUBCs were calculated assigning a value of 1,024 to all values ≥1,024 of the experimental or predicted titers. The trapezoidal rule was used for the calculation.

Inter- and intragroup comparisons of drug concentrations and bactericidal titers in urine were performed by two-way analysis of variance (treatment and phase of the trial) by repeated measures. The results were confirmed by nonparametric analysis (Mann-Whitney test). A P < 0.01 was considered statistically significant.

Modal MIC/MBC values (in micrograms/milliliter) were 0.008/0.008 for gemifloxacin and 0.015/0.06 for trovafloxacin against E. coli ATCC 25922 and 0.015/0.03 for gemifloxacin and 0.06/0.12 for trovafloxacin against S. saprophyticus ATCC 1970. When urine was the test medium, MICs and MBCs increased to 0.12/0.12 μg/ml for gemifloxacin against E. coli and S. saprophyticus and to 0.12/0.12 against E. coli and 0.25/0.5 against S. saprophyticus for trovafloxacin.

Concentrations of gemifloxacin and trovafloxacin determined by bioassay and HPLC are shown in Table 1. Concentrations of gemifloxacin and trovafloxacin determined by HPLC and microbiological assay correlated well when a linear regression (r² = 0.78 and r² = 0.80 for gemifloxacin and trovafloxacin, respectively) was used, but levels determined by bioassay were lower (without significant differences) than those measured by HPLC for gemifloxacin during the first 24 h (y = 1.3131x + 2.6802) but not for trovafloxacin (y = 0.8095x + 0.4337). Significantly higher gemifloxacin versus trovafloxacin concentrations were found over the 72-h sampling time. Due to these facts, the urine AUC determined by HPLC was higher than the AUC determined by bioassay for gemifloxacin, but no such differences were observed for trovafloxacin.

TABLE 1.

Gemifloxacin and trovafloxacin concentrations determined in urine by HPLC and by bioassay and AUC

Sample time (h)	No. of samples	Mean concn (μg/ml) ± SD of:
		Gemifloxacin		Trovafloxacin
		HPLC	Bioassay	HPLC	Bioassay
0–2	12	42.7 ± 31.1	26.5 ± 18.4	2.8 ± 1.5	2.5 ± 1.2
2–4	12	43.7 ± 13.1	28.4 ± 13.0	3.2 ± 0.8	3.6 ± 1.5
4–8	12	30.1 ± 9.8	21.7 ± 8.6	3.0 ± 0.8	3.2 ± 1.1
8–12	12	43.9 ± 22.3	29.8 ± 15.4	4.4 ± 2.0	4.7 ± 2.2
12–16	11	31.1 ± 12.0	27.2 ± 13.3	3.9 ± 1.8	4.3 ± 2.1
16–24	11	23.2 ± 10.3	14.7 ± 8.8	3.2 ± 1.4	3.0 ± 1.6
24–36	12	5.8 ± 2.5	4.9 ± 2.5	1.0 ± 0.4	1.1 ± 0.6
36–48	12	2.7 ± 1.5	1.9 ± 1.0	0.8 ± 0.5	0.6 ± 0.4
48–60	12	0.9 ± 0.9	0.9 ± 0.8	0.2 ± 0.2	0.3 ± 0.3
60–72	12	0.6 ± 0.3	0.5 ± 0.3	0.2 ± 0.1	0.1 ± 0.1
NAa (AUC [mg/ml · h])		873 ± 249	623 ± 232	106 ± 36	108 ± 45

^aNA, sample time not applicable to AUC values. 

Experimental and predicted UBTs (determined with HPLC and bioassay data) and AUBCs for gemifloxacin and trovafloxacin against E. coli and S. saprophyticus are included in Tables 2 and 3, respectively. No significant differences were found between titers predicted with HPLC and the bioassay concentrations for both quinolones against both bacteria.

TABLE 2.

Experimental and predicted (with HPLC and bioassay levels) urine bactericidal titers against E. coli and the AUBCs

Sample time (h)	Median gemifloxacin titer (range)			Median trovafloxacin titer (range)
	Experimental	HPLC	Bioassay	Experimental	HPLC	Bioassay
0	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)
0–2	≥1,024 (256–≥1,024)	4,096 (1,024–8,192)	2,048 (256–4,096)	32 (8–128)	32 (16–64)	32 (16–64)
2–4	≥1,024 (128–≥1,024)	4,096 (2,048–8,192)	2,048 (256–4,096)	32 (16–256)	32 (32–64)	32 (32–64)
4–8	≥1,024 (128–≥1,024)	2,048 (2,048–4,096)	2,048 (128–4,096)	48 (16–512)	32 (16–64)	32 (16–64)
8–12	≥1,024 (256–≥1,024)	3,072 (2,048–8,192)	2,048 (128–4,096)	64 (16–1,024)	48 (16–128)	48 (16–128)
12–16	512 (64–≥1,024)	4,096 (1,024–4,096)	2,048 (128–4,096)	32 (8–256)	32 (16–64)	64 (16–128)
16–24	512 (64–≥1,024)	2,048 (1,024–4,096)	1,024 (128–2,048)	32 (8–128)	32 (16–64)	32 (8–64)
24–36	128 (32–512)	512 (256–1,024)	512 (32–1,024)	16 (4–128)	12 (4–16)	12 (4–32)
36–48	32 (4–128)	256 (64–512)	192 (32–512)	8 (2–32)	8 (2–16)	8 (2–16)
48–60	16 (8–32)	64 (32–256)	64 (8–256)	4 (2–8)	2 (<2–8)	2 (<2–16)
60–72	8 (4–32)	64 (32–128)	64 (8–128)	2 (<2–8)	2 (<2–4)	<2 (<2–4)
NAa (AUBC)	19,532 (3,672–32,064)	32,860 (29,109–45,720)	31,663 (4,366–38,933)	1,188 (440–16,936)	1,104 (717–1,915)	1,164 (649–2,748)

^aNA, sample time not applicable to AUBC values. 

TABLE 3.

Experimental and predicted (with HPLC and bioassay levels) urine bactericidal titers against S. saprophyticus and the AUBCs

Sample time (h)	Median gemifloxacin titer (range)			Median trovafloxacin titer (range)
	Experimental	HPLC	Bioassay	Experimental	HPLC	Bioassay
0	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)	<2 (<2–<2)
0–2	256 (64–1,024)	1,024 (256–2,048)	512 (64–1,024)	8 (4–16)	16 (8–32)	16 (8–32)
2–4	256 (64–1,024)	1,024 (512–2,048)	512 (64–1,024)	8 (4–16)	16 (16–32)	16 (16–32)
4–8	256 (64–1,024)	512 (512–1,024)	512 (32–1,024)	8 (4–16)	16 (8–32)	16 (8–32)
8–12	256 (64–1,024)	768 (512–2,048)	512 (32–1,024)	16 (8–16)	24 (8–64)	24 (8–64)
12–16	128 (64–1,024)	512 (256–1,024)	512 (32–1,024)	8 (4–16)	16 (8–32)	32 (8–64)
16–24	64 (32–512)	512 (256–1,024)	256 (32–512)	8 (2–8)	16 (8–32)	16 (4–32)
24–36	48 (16–64)	128 (64–256)	128 (8–256)	4 (<2–4)	6 (2–8)	6 (2–16)
36–48	16 (4–32)	64 (16–128)	48 (8–128)	2 (<2–4)	4 (<2–8)	4 (<2–8)
48–60	4 (2–16)	16 (8–64)	12 (2–64)	<2 (<2–4)	<2 (<2–4)	<2 (<2–8)
60–72	4 (<2–8)	16 (8–32)	8 (2–32)	<2 (<2–<2)	<2 (<2–2)	<2 (<2–2)
NAa (AUBC)	5,500 (2,220–17,384)	18,397 (12,118–29,386)	14,580 (1,092–22,704)	287 (148–400)	572 (377–958)	595 (343–1,374)

^aNA, sample time not applicable to AUBC values. 

Only gemifloxacin against E. coli presented significant differences between experimental and predicted titers over the 72-h sampling period. No differences were found for trovafloxacin and E. coli. With respect to S. saprophyticus, trovafloxacin presented significant differences over the first 24 h and for gemifloxacin these occurred only at two sampling intervals (36 to 48 and 60 to 72 h), when experimental and predicted titers were compared. When AUBCs were compared, similar values were obtained for predicted HPLC and bioassay AUBCs for both antimicrobials against both bacteria, while lower values for experimental AUBCs were obtained for S. saprophyticus with both drugs and with gemifloxacin against E. coli. Significantly higher experimental bactericidal titers were obtained over 72 h for gemifloxacin versus trovafloxacin against both bacteria. This is also reflected by AUBC values expressing the same significant differences.

Reasons for possible discrepancies between predicted and experimental activities depend on the microbiological and/or pharmacokinetic parameters on which the predictions are based. With respect to pharmacokinetic parameters, gemifloxacin concentrations obtained by bioassay are clearly lower than those obtained by HPLC during the first 24 h. The discrepancy is most marked with high concentrations. This could be due to different sample processings prior to HPLC and bioassay determinations. It is possible that gemifloxacin crystals (obtained after the freeze-thaw cycle of the samples with high gemifloxacin concentrations) were redissolved prior to the HPLC determination due to agitation for 30 min at 37°C and centrifugation (Clinical Pharmacology Department, SmithKline Beecham Pharmaceuticals [personal communication]), whereas bioassay determination was performed immediately after thawing without the redissolution procedure. Crystallization was confirmed by microscopic examination of a filter used for urine filtration. On the other hand, the trovafloxacin concentrations were very similar when measured by both methods, suggesting an absence of crystallization after freezing.

With respect to the microbiological activity, urine increased the MIC and MBC values of both quinolones against the two strains used. The pharmacodynamic effect can be explored by the relationship of the AUC and the in vitro bactericidal activity (MBC). The pharmacodynamic effect of the decreased antibacterial activity by urine is much higher for gemifloxacin and E. coli (a 15-fold decrease of AUC_{0–72 h}/MBC in urine at 5,189 ± 1,934 versus AUC_{0–72 h}/MBC in broth at 77,840 ± 29,021) than for gemifloxacin and S. saprophyticus or for trovafloxacin with both strains, where the decrease is within a range of two- to fourfold range (AUC_0–72 h/MBC in urine versus AUC_0–72 h/MBC in broth: 896 ± 373 versus 1,793 ± 745 for trovafloxacin and E. coli, 215 ± 89 versus 896 ± 373 for trovafloxacin and S. saprophyticus, and 5,189 ± 1,934 versus 20,757 ± 7,739 for gemifloxacin and S. saprophyticus).

Therefore, the greater effect of urine on the AUC_0–72 h/MBC is reflected in lower experimental versus predicted titers with gemifloxacin and E. coli but not with S. saprophyticus or trovafloxacin and the two bacteria.

The lack of differences between HPLC (complete redissolution of crystallization)- and bioassay (presence of crystals)-predicted titers can be attributed to the higher specific weight of the MBC versus the antimicrobial concentrations in the method of calculation of the predicted titers (1/2ⁿ, where n is the number of dilutions for which the drug concentration remained above the MBC [3]).

From the pharmacodynamic point of view, gemifloxacin displays significantly higher ex vivo antibacterial activity than does trovafloxacin (despite the crystallization and the urine effect on its antibacterial activity), providing adequate cover (UBTs of ≥4) (4) over 72 and 48 h against the E. coli and S. saprophyticus strains, respectively.