Nontuberculosis mycobacteria infections: would there be pharmacodynamics without pharmacokinetics?

From the authors:

We appreciate the comments from H.Y. Kim and co-workers on our individual patient data meta-analysis on Mycobacterium abscessus pulmonary disease (MAB-PD) [1]. They emphasised the importance of pharmacokinetic (PK)/pharmacodynamic (PD) analysis and therapeutic drug monitoring (TDM) in the field of nontuberculous mycobacterial (NTM) disease, especially for MAB-PD. We generally agree with their perspective.

As H.Y. Kim and co-workers pointed out, most of the recommendations for MAB-PD treatment stem from empirical data and experts' opinions. In our study [1], we reported the importance of azithromycin, amikacin and imipenem in the treatment of M. abscessus subspecies abscessus. However, optimum doses of these drugs have not yet been firmly established. While the American Thoracic Society/Infectious Diseases Society of America recommended the daily use of 250 mg azithromycin [2], the British Thoracic Society suggested a dosage of up to 500 mg daily [3]. The use of amikacin has been recommended with variable dosing, from 7–10 mg·kg⁻¹ to 10–30 mg·kg⁻¹, depending on the frequency of administration, and the targeted peak concentrations (C_max) have also varied from 20–25 μg·mL⁻¹ to 35–45 μg·mL⁻¹ [4]. These variable suggestions underscore the importance of PK/PD analysis in the treatment of NTM pulmonary disease. Although two out of eight studies included in our analysis reported PK data, as H.Y. Kim and co-workers indicated, we were unable to include these PK data in our meta-analysis because other studies lacked such information.

Previously, van Ingen et al. [5] reported the PK/PD parameters in patients with Mycobacterium avium complex pulmonary disease (MAC-PD) (J. van Ingen is one of the co-authors of our study). C_max and areas under the time–concentration curve (AUC) were measured in 531 patients and the minimum inhibitory concentration (MIC) of their isolates analysed. Although these patients were being treated with currently recommended doses, the desirable target ratio of C_max/MIC or AUC/MIC for clarithromycin and amikacin were attained only in 42% and 19% of patients, respectively.

In the case of M. abscessus, incubation at sub-inhibitory concentrations of macrolides resulted in a marked increase of MIC related to inducible resistance [6]. Prolonged incubation of M. abscessus with exposure to 1× MIC of either clarithromycin or azithromycin yielded a >30-fold increase in the MIC as well as increased expression of erm(41) after 7 days of inoculation. In addition, the sub-inhibitory doses of amikacin changed the phenotypes of M. abscessus from smooth to rough colony morphology [7]. Defining optimal dosage of core drugs based on PK/PD analysis is pivotal to improving treatment outcomes of MAB-PD. However, the current regimens were suggested by results of in vitro studies and the data from several case series [2, 3]. PK/PD analysis in real-life clinical practice is somewhat tricky because of the complexity of treatment, including use of multiple drugs and toxicity issues limiting the choice and dosing of key antibiotics [2, 3]. The hollow-fibre system models using Monte Carlo simulations could be an alternative to conventional PK/PD study in humans [8].

Lee et al. [4] reported their experience using TDM for patients with MAB-PD (W-J. Koh, a co-author with Lee et al. [4], is another of the co-authors of our study). They adjusted the amikacin dose to achieve the target C_max of 55–65 μg·mL⁻¹. In week 4, 46 out of 48 patients (95.8%) achieved the target level, while transient nephrotoxicity developed in 6.3% of patients and ototoxicity in 25.0%. Although the treatment of MAB-PD necessitates the long-term use of intravenous drugs, safety and tolerability can be achieved if TDM is applied.

One pitfall in PK/PD analysis and TDM is the uncertain correlation with treatment outcomes. These stem from various reasons, including use of multiple drugs, the time- rather than concentration-dependent activity of some key antimycobacterial drugs (cefoxitin, imipenem and macrolides) [8] and the fact that rifampicin and ethambutol feature in MAC-PD regimens mainly for their synergy and to prevent macrolide resistance, not for their killing capacity [2, 3]. In MAC-PD, low plasma clarithromycin concentration was reported not to be associated with treatment outcomes [9], although subsequent study reported that a higher C_max of azithromycin was associated with favourable outcomes [10]. The proportion of patients satisfying the PD target for amikacin was 19%, which could not explain the treatment success rates for MAC-PD reaching 60% [5]. These data suggested that the clinical implication of PK/PD parameters must be interpreted with caution. Future studies such as the multinational FORMAT trial (Finding the Optimum Regimen for Mycobacterium abscessus) will include PK/PD analysis of all drugs, combined with clinical and microbiological correlations, to improve treatment outcomes of MAB-PD.

Shareable PDF