Increased Doses Lead to Higher Drug Exposures of Levofloxacin for Treatment of Tuberculosis

Fluoroquinolones, including levofloxacin, display concentration-dependent killing of Mycobacterium tuberculosis (1, 2). Typical daily levofloxacin doses (750 to 1,000 mg) do not reach high AUC/MIC (area under the concentration-time curve over 24 h in the steady state to MIC ratio) targets that have been suggested by studies of the treatment of bacterial pneumonia and murine studies of tuberculosis (TB) (3–7). High levofloxacin exposures could offer less selection of drug-resistant mutants and a shorter time to sputum culture conversion and relapse-free cure (8, 9). The Opti-Q study (ClinicalTrials registration no. NCT01918397) was a phase II, double-blinded, randomized, dose-ranging clinical trial in patients with multidrug-resistant tuberculosis (MDR TB). We compared 11, 14, 17, and 20 mg/kg/day of levofloxacin administered orally as a single daily dose 7 days per week for 26 weeks, along with an optimized background regimen of second-line TB medications, in patients in Peru and South Africa. Weight banding gave respective doses of 750, 1,000, 1,250, and 1,500 mg daily for patients who weighed ≥60 kg and, due to a minimum dose floor, 750, 750, 1,000, and 1,250 mg daily for patients who weighed <60 kg. Here, we report the noncompartmental pharmacokinetic (PK) results of this comparison.

This study was reviewed and approved by the institutional review board of each participating institution. The study included consenting adults with newly diagnosed, previously untreated, smear-positive (≥2+) pulmonary MDR TB. We used line probe results (MTBDRplus and MTBDRsl; Hain, Nehren, Germany) to screen patients for eligibility pending phenotypic drug susceptibility testing results. Those showing isoniazid and rifampin resistance and fluoroquinolone susceptibility were eligible. Other inclusion criteria included known HIV status (regardless of result and therapy), a weight of ≥40 kg, and a Karnofsky Performance Status score of >60. Full details of eligibility criteria and trial design are available in the published protocol (10). All study treatment doses were directly observed. The optimized background regimen (OBR), comprised of other second-line drugs without a fluoroquinolone, was selected by local investigators in order to conform with local standards of care and guidelines. Levofloxacin 250-mg capsules and matching placebo 250-mg capsules were provided by Macleods Pharmaceuticals and combined in a dose package by the unblinded pharmacist at each site. Each participant received 6 tablets, but investigators, clinicians, and participants were blinded to the dose of levofloxacin. All treatment was ambulatory and supervised by study staff. Study participants were advised to avoid aluminum- and magnesium-containing antacids within several hours of each levofloxacin dosing. Food intake was recorded but not restricted. After 14 to 28 total and at least 3 consecutive daily doses of levofloxacin, plasma samples were collected at 0, 1, 2, 4, 8, 12, and 24 h postdose. Samples were shipped frozen and stored at −80°C until assayed at the University of Florida using a validated high-performance liquid chromatography (HPLC) assay with fluorescence detection. The plasma standard curve for levofloxacin ranged from 0.20 to 15 μg/ml; overall precision was 0.58% to 4.09% (coefficient of variation); quality control sample precision was 2.88% to 3.79%. Phoenix v7.0 (Certara LP, Princeton, NJ) was used for noncompartmental analysis and JMP 10 (SAS Institute, Cary, NC) for Y by X nonparametric statistical analysis (Kruskal-Wallis test, chi-square test) or linear regression (adjusted R², analysis of variance [ANOVA]). Pairwise comparisons were made using Tukey's honestly significant difference (HSD) test and a comparison of each pair using Student's t test.

Among the 111 participants randomized, 101 participants had evaluable pharmacokinetic data. There were 22, 26, 26, and 27 patients evaluable in the 11-, 14-, 17-, and 20-mg/kg/day levofloxacin dose groups, respectively. Participant characteristics and the levofloxacin doses received are presented in Table 1. Pharmacokinetic results for each of the dosing groups are presented in Tables 2 and 3. The maximum plasma concentration (C_max) ranged from a low of 5.82 to a high of 35.42 μg/ml (roughly 6-fold range). For comparison, the normal C_max range for 750 to 1,000-mg doses is 8 to 12 μg/ml (11). Within each dosing group, the C_max varied approximately 3-fold (Fig. 1). Four of 36 patients (10%) who received a 750-mg dose (using an 11 or 14-mg/kg dose) achieved a C_max value below 8 μg/ml, the low end of the normal range (11). No patients who received 1,000 mg levofloxacin or more had a C_max value below 8 μg/ml. Due to weight banding and the 750-mg minimum dose, 14 (54%) patients in the 14-mg/kg group received the same dose as patients in the 11-mg/kg group (Table 1).

When the area under the concentration-time curve from 0 to 24 h (AUC₀₋₂₄) was evaluated by the milligram dose administered, it increased proportionally. Within each mg/kg dosing group, the AUC₀₋₂₄ varied approximately 3-fold (Fig. 2). The median T_max was 2 h across all groups, and median half-lives (t_1/2) were 6.1 to 6.7 h. Thus, increasing levofloxacin doses did not change the apparent rates of absorption or elimination. In univariate analyses, females had higher C_maxs and higher AUC₀₋₂₄s, despite having shorter t_1/2s, than the males (Table 4), although there were slightly more female patients in the two higher-dose groups combined (24 [45%] of 53) than in the two lower-dose groups combined (17 [35%] of 48). HIV-positive patients had slightly higher C_maxs, higher AUC_0–24s, and longer t_1/2s than the HIV-negative patients (Table 4). Across all ages, older patients had higher AUC_0–24s and longer t_1/2s than the younger patients (Table 4). In a multivariate analysis of these three key covariates, sex remained significant for C_max, while sex, HIV status, and age all remained significant covariates for AUC_0–24 and t_1/2. The independent data safety and monitoring board (DSMB) met every 6 months throughout the study, which was not interrupted at any point due to safety concerns; a full analysis of the safety data is under way.

These results demonstrate that increased doses of levofloxacin in the presence of a background MDR-TB regimen lead to increased levofloxacin C_maxs and AUC_0–24s. Thus, patients with low serum levofloxacin concentrations can be expected to respond to increased dosing. Moreover, if increased serum concentrations of levofloxacin are associated with increased clinical efficacy without dose-limiting toxicity, increased dosing may improve the proportion of favorable treatment responses. If proven safe and efficacious, these pharmacokinetic data support further exploration of high-dose levofloxacin in MDR-TB regimens.

We thank all the study participants who contributed their time to this study. We are grateful to the local TB program staff who assisted in the clinical management of study participants. Finally, we thank the CDC Tuberculosis Trials Consortium members and staff for encouragement and thoughtful advice throughout the study.

This work was supported by the U.S. National Institutes of Health NIAID grant U01 AI100805, CFAR grant P30 AI042853, the U.S. CDC DTBE TBTC, and the U.S. National Institutes of Health NIAID grant 1K24 AI114444. Additional support was provided by the Providence/Boston Center for AIDS Research (P30AI042853) and the Boston University/Rutgers Tuberculosis Research Unit (U19AI111276).