Trovafloxacin Concentrations in Airway Fluids of Patients with Severe Community-Acquired Pneumonia

Abstract

The penetration of trovafloxacin (TVA), 200 mg once daily, into the airways of 17 patients with severe pneumonia was studied. The mean (standard deviations are given in parentheses) steady-state TVA concentrations, 2 h after the last intake, were 3.1 (0.3) mg/liter in induced sputum (n = 8), 3.2 (1.1) mg/liter in bronchial secretions (n = 9), 3.2 (0.9) mg/liter in bronchoalveolar lavage fluid (n = 10), and 4.9 (1.4) mg/liter in epithelial lining fluid (n = 11).

Trovafloxacin (TVA) belongs to a new category of quinolone antibiotics whose members have in common enhanced activity against gram-positive bacteria (5, 7, 12), making this class of antibiotics the mainstay of therapy in community-acquired respiratory tract infections (4). As both bronchial secretions and bronchial mucosa concentrations of antibiotics are considered to be predictive of the outcome of therapy of lower respiratory tract infections (3), we assessed the concentrations of TVA in serum and in airway fluids of patients with clinically and radiologically proven pneumonia.

Seventeen consecutive patients (>18 years old) with severe acute community-acquired pneumonia, recruited during the winter of 1997 to 1998, were included. Diagnosis was based on the presence of a new infiltrate on chest radiography, fever greater than 38°C, and at least one of the typical clinical signs (cough, dyspnea, chills, sputum production, or chest pain) or a white blood cell count of 10,000 to 30,000/μl. Common exclusion criteria were applied. The study was approved by the hospital Ethics Committee, and all patients gave written informed consent.

Patients received 200 mg of TVA orally once daily for 10 days. Treatment started as soon as study admission procedures were completed but before the results of culture or serology were available. Acceptable culture material included expectorated sputum (provided that >25 neutrophils and <10 epithelial cells were found per low power field), material from bronchoscopic aspiration, pleural fluid, and blood. Serum samples were obtained and frozen on admission and at the end of treatment for serological testing for Legionella pneumophila, Mycoplasma pneumoniae, Chlamydia spp., and Coxiella burnetii. Only pathogens that were possibly relevant for respiratory tract infection were considered for microbiological documentation of the study. Atypical pneumonia was considered to be documented in the case of a fourfold rise of antibody titers or a single titer of ≥8 for immunoglobulin M M. pneumoniae- or Chlamydia pneumoniae-specific immunofluorescent antibodies.

To determine TVA concentrations, sputum was induced on the fourth day and processed as previously described (10). Fiber bronchoscopy was performed on the fifth day. A standard bronchoalveolar lavage (BAL) was performed by using 200 ml of prewarmed 0.9% saline, aliquoted into four 50-ml aliquots. The dwell time was approximately 30 s. The aspirate from trachea and bronchi before instilling the isotonic saline was labeled bronchial secretions. The first aspirate (BAL fluid) was kept separately; aspirates from the remaining three aliquots were pooled and used to calculate epithelial lining fluid (ELF) volume. In each case, the samples obtained through bronchoscopy were centrifuged immediately at 400 × g for 5 min and the supernatant was separated from the cells without delay and aliquoted to allow measurements of urea and TVA concentrations. The mean intervals between administration of TVA and sampling of specimen (induced sputum or BAL fluid) were 120 and 135 min, respectively. TVA concentrations in serum were determined according to the method of Teng et al. (13). The method was slightly modified for the determination of TVA in BAL fluid and sputum samples. The samples (200 μl) were determined by high-pressure liquid chromatography and UV detection (275 nm). The separations were performed on a 3.9- by 150-mm Symmetry C₁₈ column (5 μm) with a 3.9- by 20-mm Symmetry Sentry C₁₈ guard column (5 μm). The mobile phase was a mixture of acetonitrile–0.015 M NaH₂PO₄–acetic acid–tetrahydrofuran (18:80:1:1). The retention times of the internal standard (a methyl derivative of TVA) and TVA were 9.5 and 7.5 min, respectively. The lower limit of quantification for TVA was 50 ng/ml (using 200 μl of fluid). The within-run and between-run coefficients of variation were less than 2% in plasma (500, 2,000, and 4,000 ng/ml; n = 6) and less than 3% in BAL fluid (250 and 375 ng/ml; n = 6). The within-run and between-run accuracy in plasma and in BAL fluid ranged from 100 to 105%. Six quality control samples together with a calibration curve were analyzed with each run of patient samples. The assay was linear over the range of 50 to 7,000 ng/ml. Using the dilution method described by Renard et al. (11), the concentration of TVA in ELF was determined as follows. ELF antibiotic concentration = (ACL × BL)/UL, where ACL is the antibiotic concentration in lavage (milligrams per liter), UL is the urea concentration in lavage (millimoles per liter), and BL is the blood urea concentration (millimoles per liter). TVA concentrations in bronchial secretions and BAL fluid were determined in a similar way. The concentration of TVA in sputum was calculated by taking into account the dilution factor. The concentration of urea in bronchial secretions and ELF was determined by using a modified Roche diagnostic kit on a Hitachi 747 autoanalyzer.

Statistical analyses were performed with an SAS software package (SAS Institute Inc., Cary, N.C.). One-way analysis of variance or the Kruskal-Wallis test was used as appropriate. A P value of <0.05 was considered to be significant.

Patient characteristics are given in Table 1. The demographic and baseline parameters were comparable between patients with bacterial or atypical pneumonia. The differentiation between bacterial and atypical pneumonia was made retrospectively. After at least 4 days of treatment, no pathogens could be isolated anymore in eight responding patients with presumed bacterial pneumonia, whereas Pseudomonas aeruginosa was isolated in one patient considered to be improved. Nocardia asteroides was grown from the bronchial aspirate of a nonresponding patient. Antibodies to M. pneumoniae were detected in four patients, and antibodies to Chlamydia spp. were detected in three patients. All seven patients were cured.

TABLE 1.

Clinical characteristics and response of the 17 patients, according to type of pneumonia

Clinical characteristics	Bacterial pneumonia	Atypical pneumonia
n	10	7
Age (yr ± SD)	70 ± 3	55 ± 12
Sex (male/female)	9/1	2/5
Fever (°C ± SD)	38.2 ± 1.1	38.3 ± 1.1
Leukocytosis (103/μl ± SD)	14 ± 3	13 ± 6
Clinical efficacy
Cure	8	7
Improvement	1	0
Relapse	0	0
Failure	1	0

In all evaluable patients, similar TVA concentrations in serum, sputum, bronchial secretions, BAL fluid, and ELF were detected (Fig. 1). The mean (standard deviations are given in parentheses) TVA concentration in induced sputum (n = 8) determined on the fourth day of treatment was 3.1 (0.3) mg/liter. The TVA concentration in serum was 3.9 (0.2) mg/liter. TVA concentrations in fiberoptically obtained bronchial secretions (n = 9), BAL fluid (n = 10), and ELF (n = 11) were determined on day 5. Mean TVA concentrations were 3.2 (0.9), 3.2 (1.1), and 4.9 (1.4) mg/liter, respectively. The mean peak serum concentration was 3.4 (0.5) mg/liter.

FIG. 1.

Individual data of TVA concentrations in serum on day 4 (s4) and day 5 (s5), in sputum (sp), bronchial secretions (bs), BAL fluid, and ELF. Horizontal bars represent the means. No difference was observed between mean peak serum concentrations and the concentration of TVA in the sampled respiratory tract compartments.

Several important conclusions emerge from our study, which evaluated TVA concentrations in sputum, bronchial secretions, BAL fluid, and ELF from patients with pneumonia. First, high concentrations were found in all sampled compartments of the respiratory tract. Second, peak and through concentrations in serum exceeded reported MICs (1, 6, 9) for common respiratory tract pathogens, and lastly, concentrations of TVA in induced sputum are not significantly different from the concentrations in other respiratory tract compartments.

The demonstration of persistent high TVA concentrations in secretions of the respiratory tract confirms and extends, in a larger group of patients, previous findings by Andrews et al. (2). These investigators used a different method to determine the concentration of TVA. In our study, concentrations were measured in steady-state conditions after at least 4 days of treatment. When comparing TVA concentrations in ELF, bronchial secretions, BAL fluid, and sputum, no significant difference was noted between these compartments. Furthermore, as both bronchial secretions and bronchial mucosa concentrations of antibiotics are considered to be predictive of the outcome of therapy of lower respiratory tract infections (3), the demonstration of comparable concentrations in sputum may allow for the use of induced sputum as a tool to monitor the outcome of the treatment. As sputum induction is a safe and easy reproducible way to obtain secretions from the lower airways, both in asthmatics (8) and in chronic obstructive pulmonary disease patients (10), our data support the use of this technique in patients with respiratory tract infections as well and suggest that measurement of TVA concentrations in sputum may help to predict the efficacy of antibiotics. In conclusion, we demonstrate that high concentrations of TVA in serum and respiratory tract secretions are obtained in patients with severe respiratory tract infections. Our data demonstrate that determination of TVA in selected sputum plugs is feasible and suggest that this method might be studied prospectively as a possible parameter to predict clinical efficacy.