Moxifloxacin in multidrug-resistant tuberculosis: is there any indication for therapeutic drug monitoring?

To the Editors:

Fluoroquinolones are rapidly emerging as important drugs in the treatment of tuberculosis (TB) worldwide [1]. In drug-susceptible TB, their use is currently under investigation and, according to the American Thoracic Society/Centers for Disease Control guidelines, fluoroquinolones are indicated only in patients receiving the conventional regimen who present severe adverse reactions [2]. However, in drug-resistant (DR) and, more specifically, in multidrug-resistant (MDR)-TB, the role of fluoroquinolones is much better established. Later-generation fluoroquinolones are included even in extensively drug resistant (XDR)-TB regimens since they may have some efficacy against ofloxacin-resistant strains [3].

The area under the concentration–time curve up to 24 h post-dosage (AUC₂₄) is generally considered as the best predictor of fluoroquinolone efficacy [1]. Among fluoroquinolones, moxifloxacin is considered the most bactericidal, with potency comparable to that of isoniazid, and also seems to have some sterilising activity [1]. The maximal concentration (C_max) of moxifloxacin exceeds mutant prevention concentration [1] and the currently recommended dose of 400 mg is likely to suppress the emergence of resistance in 60% of patients [4]. However, pharmacokinetic data on moxifloxacin in patients with TB are scarce, especially in the setting of an MDR-TB regimen.

In this report of a pilot prospective study, we present preliminary data on pharmacokinetic parameters in patients with MDR- or XDR-TB receiving second-line treatment. Patients were receiving moxifloxacin (Avelox®; Bayer, Leverkusen, Germany) p.o. for ≥4 days, in order to achieve a steady state, as part of their anti-TB treatment, which was based on drug susceptibility test results. Quinolone resistance was determined by ofloxacin susceptibility testing. All patients gave informed written consent and the study was approved by the hospital’s ethical committee.

Plasma samples were collected via a peripheral venous catheter immediately before and 1, 1.5, 2, 3, 4, 6, 9, 12 and 24 h after the moxifloxacin dose. Samples were stored at -20°C until they were analysed. Moxifloxacin concentration was determined by high-performance liquid chromatography with fluorescence detection as previously described [5]. The AUC₂₄ was calculated according to the trapezoidal rule.

Seven patients (five males and two females, mean±sdage 40.1±15.7 yrs) were included in the study. All patients were HIV negative, suffered from extensive pulmonary TB, and had normal renal and hepatic function. Patients received 400 mg of moxifloxacin daily, a mean dose of 5.5 mg·kg⁻¹. None of the patients was on treatment with antacids, sucralfate, drugs containing metal cations, multivitamins containing iron or zinc, nonsteroidal anti-inflammatory drugs, or class IA and III antiarrhythmics. Moxifloxacin was well tolerated and no adverse effects were reported. The mean period of treatment until sample collection was 244.1±257.3 days. Patients’ characteristics, comorbidities, anti-TB drugs, pharmacokinetic parameters and outcomes are presented in table 1. Mean C_max was 4.59±2.06 mg·L⁻¹ and was reached after 2.36±0.56 h (T_max). Mean AUC₂₄ was 37.96±16.52 mg·h·L⁻¹. Both C_max and AUC₂₄ showed a high variability.

To our knowledge, this is the first pharmacokinetic analysis of moxifloxacin in a series of patients with MDR-TB. Concerning C_max and AUC₂₄, our results are in agreement with previous reports on moxifloxacin pharmacokinetics in healthy volunteers [6]. In addition, Peloquin et al. [7] and Nijiland et al. [8], in two interesting studies administered moxifloxacin in patients with drug-susceptible TB either before or after the end of standard treatment, observed similar results. However, in the only study evaluating pharmacokinetic parameters in the setting of an anti-TB regimen, significantly lower values were detected (C_max 2.5 mg·L⁻¹, AUC₂₄ 24.8 mg·h·L⁻¹) [9]. The reasons for this discrepancy are not clear. One possible explanation might be the co-administration of rifampicin in 10 out of 16 patients for whom pharmacokinetic analysis was available, although the moxifloxacin AUC₂₄ was not statistically significantly different between patients who received rifampicin and those who did not [9]. As shown by Nijland et al. [8], rifampicin reduces moxifloxacin C_max and AUC₂₄ by ∼30%. The percentage of 30% corresponds to the difference between our results and those of Pranger et al. [9]. Rifampicin, like all absorbable rifamycins, is a well-known inducer of cytochrome P450 isoenzymes, and induces phase II metabolic processes of glucuronidation and sulfation. Since cytochrome P450 is not involved in moxifloxacin metabolism, the rifampicin–moxifloxacin interaction is expected to result from the induction of phase II metabolism. It is noteworthy that three of our patients were receiving rifabutin. To our knowledge, no trials examining the effect of rifabutin on moxifloxacin metabolism have been conducted. However, the possibility of an interaction between these two drugs cannot be excluded, although, according to our results, it would probably be weaker than the rifampicin–moxifloxacin interaction. It is known that rifabutin is a less potent enzyme inducer than rifampicin. Isoniazid, which was also administered in the study by Nijiland et al. [8], is only known to affect cytochrome P450 metabolism and, thus, it is not expected to alter moxifloxacin levels. Therefore, considering moxifloxacin levels, patients with MDR-TB receiving moxifloxacin (but not rifampicin) have an advantage in comparison with patients with DR-TB receiving both drugs. Interactions between moxifloxacin and pyrazinamide or second-line agents have not been clearly established in the literature.

In contradiction to previous data, in the present study, mean T_max was more than double the value usually observed. This discrepancy may be attributed to co-administered drugs the impact of which on the pharmacokinetics of moxifloxacin has not been investigated yet. Interestingly, T_max reached its higher value (3 h) in the two patients with the lowest C_max and AUC₂₄.

In accordance with the study by Pranger et al. [9], a high variability of C_max and AUC₂₄ values was observed. As shown in table 1, two groups of patients can be recognized, those with a C_max of ∼6 mg·L⁻¹ and an AUC₂₄>50 mg·h·L⁻¹, and those with significantly lower values. In the first group, the AUC₂₄/minimum inhibitory concentration (MIC) ratio would exceed 100 for an MIC of 0.5 mg·L⁻¹. These differences are hard to explain since patients were repeatedly instructed not to use medications with known effects on moxifloxacin metabolism. Additionally, none of the patients was underweight at the time of the pharmacokinetic analysis or during the whole anti-TB treatment, with the exception of patient 7, a patient with cystic fibrosis. In the case of patient 5, low values may be explained by poor compliance. The case of patient 4 with ofloxacin-resistant TB is more intriguing, since there were no compliance issues and low AUC₂₄ may actually have deprived this patient of a possible cure. Surprisingly, the clinical condition of this patient remained exceptionally good despite treatment failure. However, given the complexity of the second-line regimens, it is difficult to attribute this treatment failure strictly to the low levels of moxifloxacin. It has been suggested that moxifloxacin may be effective in ofloxacin-resistant TB when MIC is ≤2 mg·L⁻¹ [10]. Therefore, in agreement with Pranger et al. [9], we conclude that the standard dose of 400 mg of moxifloxacin is not sufficient for all patients and this is especially true for XDR-TB cases.

In this setting, monitoring of therapeutic concentrations of moxifloxacin, particularly in patients with XDR-TB, could have a significant impact on clinical decisions. However, it is noteworthy that the pharmacodynamic target for Mycobacterium tuberculosis has not been defined, i.e. it is not clear whether the AUC₂₄/MIC ratio should exceed 40 (as in the case of Gram-positive bacteria) or 100 (as is generally indicated in the case of Gram-negative bacterial infections). Given the complexity of anti-TB regimens and the variability of moxifloxacin concentrations among patients receiving the same regimen, it is imperative that specific pharmacodynamic targets should be determined through properly designed clinical trials and, hence, appropriate adjustments of dosage may be attempted in clinical practice.

• ©ERS 2012