University of Groningen Optimizing levofloxacin dose in the treatment of multidrug-resistant tuberculosis Ghimire, Samiksha

University of Groningen

Optimizing levofloxacin dose in the treatment of multidrug-resistant tuberculosis

Ghimire, Samiksha

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Ghimire, S. (2019). Optimizing levofloxacin dose in the treatment of multidrug-resistant tuberculosis: An integrated PK/PD approach. University of Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

7

Discussion

and future perspectives

7

antimicrobial therapy has two pillars; pharmacokinetics and pharma-codynamics. Pharmacokinetics (PK) describes how the human body handles a drug; it describes the processes relating to drug absorption, distribution, metabolism and excretion (1). Pharmacodynamics (PD) describes the effect of drug on the human body; in antimicrobial drugs, it describes the interaction with the target site of the causative organism (efficacy) aa well as the impact on the patient (toxicity) (1). MIC is the Minimal Inhibitory Concentration of a drug necessary in a in vitro system that inhibits the growth of the micro-organism (1). Integration of PK and PD (PK/PD) links a PK parameter together with MIC such as AUC0–24/MIC, Cmax/MIC, Cmin/MIC or %T>MIC to the clinical

effect (1). Then, the optimal dose and dosing frequency is determined based on the probability of target attainment for maximum Mtb kill or prevention of resistance. A good agreement exists in PK/PD parameter between pre-clinical (hollow fiber infection model studies and animal studies) and clinical studies (1, 2).

The efficacy of levofloxacin (Lfx) is predicted by its ratio of area under the 24-h concentration time curve (AUC0–24) and minimum inhibitory concentration (MIC) (3). Until recently, a target value of >100–125 based on infection by Gram-negative bacteria and >40 based on infection by Gram-positive bacteria on in vitro and animal models, and human data were established (4, 5). In the absence of data on TB bacteria, AUC0–24/ MIC >100–125 was generalized for under-standing Lfx dose-concentration-response relationship in TB patients (6–9). Available pharmacokinetic data from clinical studies showed that at least 25 % of the patients on standard once daily Lfx dosages of 750–1000 mg dosing did not achieve the desired AUC0–24, Cmax

and AUC0–24/MIC ratio (Chapter 2) (10). The sub-therapeutic drug levels and pharmacokinetic variability could be crucial in fueling the emergence of resistance. This became evident in a cohort study where 11.2 % of the patients without any baseline resistance developed ac-quired FQs resistance on treatment with a standard dose (11, 12). The same phenomenon has also been observed for key first line drugs like isoniazid and rifampicin (2). Lfx is the backbone of the combination regimen in use to treat MDR-TB due to its excellent efficacy and good safety profile (13). An individual patient data meta-analysis showed

G en era l di sc us sio n a nd f ut ur e p er sp ec tiv es

that presence of Lfx in the regimen was positively associated with treatment success, with an adjusted risk difference of 0.15 (95 % CI 0.13–0.18), compared to failure or relapse (14). Therefore, losing this drug to under dosing is unaffordable.

To be able to define the most appropriate Lfx dose in TB patients, following knowledge gap needed to be bridged by three major areas of research. First, the PK/PD target for optimal exposure needed to be established (10). Second, optimal dose finding and evaluation of the possibility of individual dose optimization through therapeutic drug monitoring (TDM) should be investigated (10). Third, there was the need to redefine susceptibility breakpoints from the conventional binary system that separated susceptible from resistant strains based on critical concentration to S-I-R system (15–17).

Lfx PK/PD target for optimal exposure and defining optimal dosing

Given as a monotherapy, the hollow fiber model on tuberculosis recently established an Lfx target of 146 for maximum bacterial kill (EC80) and 360 for the prevention of acquired drug resistance (3). Here we see, that the PK/PD driver for prevention of acquired drug resistance is not sim-ilar to that of bacterial kill. In Chapter 6a, we explored this new target in a prospective clinical study that incorporated all three components: exposure, efficacy and treatment outcomes. We found that 30 % of the patients did not attain AUC0–24/MIC >146 and only 50 % of the patients with elevated MICs (1 mg/L) met the target. The lower probability of target attainment in patients infected by clinical isolates with elevated MICs (85 % at MIC of 0.5 mg/L and 30 % at 1 mg/L) reflects the prob-lems with the currently used 11–14 mg/kg dosing scheme and need for dose increment, given the fact that benefits of efficacy outweigh the risk of toxicity. Regarding toxicity, the use of Lfx has been associated with several warnings from FDA and EMA on disabling and potentially permanent side effects involving tendons, muscles, joints, nerves, and central nervous system (18, 19). Additionally, Lfx should be used with caution in patients with diminished renal function since it is primarily excreted by kidneys in unchanged form (>80 %) (20). Furthermore, the concomitant use of corticosteroids in these patients is known to aggravate its serious adverse effects (20).

7

phase II dose-finding trial (OptiQ trial, NCT 01918397) comparing four weight-based regimens of Lfx 11, 14, 17 and 20 mg/kg/day in MDR-TB patients (21). Initial pharmacokinetic data showed that higher doses from 17 to 20 mg/kg/day (equivalent to an actual daily dose of 1250 and 1500 mg) showed a three-fold proportional increase in area under the concentration time curve (AUC0–24) compared to currently used 750–1000 mg once daily dosing, within each mg/kg dosing group (22). The OptiQ trial is further investigating the linearity between the AUC/MIC ratios and primary end points (time to stable culture con-version). If the final results show improved efficacy on higher doses, without or with an acceptable increased risk of toxicity, Lfx dosages would need to be increased to 17–20 mg/kg (21, 22).

Role of TDM

In Chapter 6a, high inter-individual variabilities (CV% min, max: 19.13 %, 67.28 %) in Lfx plasma concentrations were observed on stan-dard dosages. The concentrations were quantified by a developed and validated bio-analytical method using liquid-chromatography tan-dem-mass spectrometry (LC-MS/MS) (Chapter 5). Due to improved selectivity and specificity of this LC-MS/MS method, multiple anti-TB drugs can be analyzed at the same time. This has minimized the re-quired sample volume, reduced sample preparation time and shortened run times. For implementing TDM in programmatic settings, a single assay able to analyze multiple anti-TB drugs will be essential. Kim et al. published a method for simultaneous analysis of 20 MDR-TB drugs using LC/MS-MS (23) and Kuhlin et al. have provided an overview of development of multi-analyte assay for TDM (24).

The inter-individual pharmacokinetic variability in humans is the result of genetic diversity that leads to the polymorphisms of trans-porters and drug metabolizing enzymes, differences in gut function, and variable hepatic and renal function which alter the rate of absorp-tion, bioavailability, distribuabsorp-tion, metabolism and excretion of drugs (25, 26). In addition, lean and fat body mass, and body weight/body mass index, all contribute to the variability (25). This explains why different individuals achieve a range of concentrations on the same standard dosages. In addition, predisposing risk factors can influence

G en era l di sc us sio n a nd f ut ur e p er sp ec tiv es

drug exposure. Some risk factors such as HIV, malnutrition, overweight or obesity, short bowel syndrome, chronic diarrhoea causes drug mal-absorption (with risk of acquired drug resistance), whereas diabetes mellitus, with or without gastroparesis causes prolonged exposure which might aggravate the toxic effects of drugs (27–29). Impaired renal and metabolic clearance may also cause excessive drug exposure. A certain threshold concentration above the MIC must be achieved by a drug to effectively kill the bug. The exposure, however, should not be above the therapeutic window to avoid increased risk of toxicity. Therefore, predicting desirable concentrations with the given standard

or mg/kg doses is difficult due to the existence of multiple sources of variability (2). This highlights the importance of drug concentration measurements and the crucial role of TDM guided dosing (1, 30–33). TDM has the potential to address too low or too high exposure by measuring the concentrations of drugs in serum/plasma of individual patients (30). The serum/plasma levels are however, not more than a proxy for the drug effect. The true effect is determined by the drug concentration at the target site (epithelial lining fluid, bone, pleural, or the core of cavitary lesions). Conte et al. studied levofloxacin con-centrations in plasma and epithelial lining fluid (ELF) in four healthy volunteers who received 750–1000 mg once daily. Levofloxacin con-centrations were significantly higher in ELF than in plasma at all time points. To date, there is a lack of data on penetration of TB drugs in ELF in TB patients. One study evaluated the penetration of Lfx in cavitary lesions (34). Encouragingly, Lfx showed excellent penetration in the cavitary lesions of chronic MDR-TB patients with a median cav-ity-to-serum ratio of 1.33 (range, 0.63–2.36) (34). Nevertheless, there was a wide variability in cavity-to-serum ratios between individuals (34). Taking the above considerations into account, plasma/serum concentration as predictor of efficacy already introduces one level of variability. This could be problematic, as Dheda and colleagues showed that resistance was driven by sub-therapeutic drug concentrations in different anatomical micro-compartments in (resected) human lungs (2). For moxifloxacin, infecting Mtb sub-populations had a variable MIC distribution in the same patient depending on the biopsy site (normal lung tissue, peri-fibriotic margin of cavity, center of granuloma, edge of necrotic cavity, 2 ml cavity caseous material, proximal airway

7

in MIC by biopsy site in the same individual reflects that having an adequate drug concentration in the blood alone is not enough (2). The drug penetration and adequate exposure into the TB lesion is more likely to determine the success or failure of therapy (2). Therefore, TDM that combines plasma/serum concentrations with MIC along with good clinical care could enhance the probability of treatment success in MDR-TB patients (35).

While TDM is a routine procedure in few treatment centers, it has yet to be adopted in the programmatic settings because of perceived constraints, despite its unprecedented role in improving treatment outcomes based on observational studies (36, 37). TDM using conven-tional venous sampling has its limitations in resource limited settings. The need of skilled manpower for venipuncture, cooling conditions for storage and transportation of venous blood samples, risk of transmis-sion of blood borne infection such as Hepatitis and HIV, and advanced bio-analytical methods for drug quantification are some of the hurdles (38). These hurdles could be overcome by using alternative and limited sampling strategies (38, 39). For Lfx, limited sampling strategies based on two sampling time-points is established. The Bayesian approach can adequately estimate AUC0–24 using samples collected at 0 h and 5 h post dose, whereas multiple linear regression requires sampling at 0 h and 4 h (39). To date, however, no prospective randomized trials have been performed comparing the superiority/non-inferiority of TDM guided treatment to non-TDM guided standardized treatment, neither have there been studies on cost-effectiveness. Diel and colleagues found the average total treatment cost per patient (direct and indirect) of €10,282 for DS-TB, €57,213 for MDR-TB and €170,744 for XDR-TB across 18 countries in Europe including the Netherlands (40). Direct costs included costs for medication, laboratory work, hospitalization and outpatient visits whereas, indirect costs represented productivity loss at work due to TB induced sickness (40). In the US, direct TB treat-ment costs per patient averaged $17,000 for non-MDR-TB, $134,000 for MDR-TB and $430,000 for XDR-TB (41). Therefore, prevention of one case of acquired drug resistance/100 patients with MDR-TB by performing TDM of FQ will certainly lighten the economic burden posed by TB disease (42).

G en era l di sc us sio n a nd f ut ur e p er sp ec tiv es

In Chapter 3, we have proposed tools and a strategy for implementing

TDM in the World Health Organization (WHO) three tier pyramid of essential TB diagnostics in a similar way TB diagnostic have been placed. The use of alternative sampling strategies (dried blood spots (DBS) and saliva) at a local and regional level, and conventional venous sampling or DBS at a regional/national level could make TDM a possibility in TB treatment (38). Nevertheless, alternative sampling strategies have their caveats. DBS sampling has stability issues for some drugs, and other drugs like amikacin do not penetrate into saliva (43). However, saliva could be a useful non-invasive matrix for drugs that have good penetration and less variable S/P ratios (44). In Chapter 6b, we found that despite a good Lfx penetration in saliva, the high variability in in-dividual saliva-to-plasma (S/P) ratios limits the use of saliva as a valid substitute for plasma. The variable S/P ratios are sensitive to the changes in salivary pH, plasma pH, drug pKa, free fraction of drug in plasma and salivary flow rate. It might not be feasible to measure these variables in clinical circumstances. For Lfx, the use of saliva might be a useful tool to effectively screen patients eligible for TDM for example using a portable spectrophotometer device at community level (45). Before using saliva, it is imperative to establish the plasma saliva correlation for individual drugs, and develop and validate semi-quantitative/quantitative chro-matographic- or spectrophotometric methods. We might be compelled to question the accuracy of TDM using semi-quantitative testing and advocate the use of only robust, advanced, high performance liquid chromatography methods using venous blood for decision making. This school of thought is rather conservative and might miss the feasible

opportunity of determining concentrations of key drugs that are out of range in TB patients. To ensure the quality and reproducibility of labora-tory results generated for TDM, internal and external quality assessments should be performed by establishing strong networking between TDM centers at the different levels of health care in a country. The LC-MS/ MS centers at the national level at a national reference laboratory could evaluate the TDM results from regional centers once every three-four months, while the regional centers could train and qualify healthcare workers for performing TDM at the community level. Externally, the national reference laboratory could be quality assured and accredited by WHO supranational TDM reference laboratories once a year.

7

rifampicin and pyrazinamide, the standard dose is inadequate (46–48). The same accounts for fluoroquinolones (49). As a result of under-

dosing, resistance has amplified (2). Furthermore, treatment failure in MDR-TB patients does not leave them with alternative retreatment options (35). While drug concentrations alone are not a substitute for clinical judgement, they could, nevertheless, be a better surrogate for estimating drug effects. TDM could further help in identifying patients with low drug exposure, who are also at a risk of acquisition of drug resistance and treatment failure. (50). In many ways, our TDM prop-osition also mirrors how TB diagnostics are placed. In TB endemic settings, national level laboratories facilitate culture in solid/liquid medium and DST by phenotypic and genotypic methods. The local and regional levels still utilize sputum smear microscopy as a primary method to diagnose TB, only some centers have the new, rapid, nucleic acid amplification test, Xpert MTB/RIF (38).

Redefining susceptibility break points for fluoroquinolones

The gold standard method for phenotypic drug susceptibility test-ing is based on the culture of Mtb in solid or liquid media (51). The Löwenstein Jensen and Middlebrook 7H10/7H11 are examples of solid media, whereas BACTEC™ Mycobacterial Growth Indicator Tube™ 960 is a liquid medium (51). Phenotypic DST utilizes epidemiological cut off values (ECOFF) also known as “critical concentrations (CCs)” of anti-TB drugs to determine the susceptibility or resistance of cul-tured Mtb. The critical concentration (CC) is defined as the lowest concentration of an anti-TB agent in vitro that will inhibit the growth of 99 % of phenotypically wild type strains of Mtb (51). The critical concentration represents highest MIC that shows phenotypical drug susceptibility and is determined by the proportion method with 1 % as the critical proportion (16, 52).

Previously used critical concentrations for both moxifloxacin and Lfx were extrapolated from ofloxacin. The newly defined CC for Lfx is 2 mg/L in Lowenstein Jensen and 1.0 mg/L in both Middlebrook 7H10 and MGIT (51). This CC based approach distinguishes suscep-tible isolates from resistant, assuming any increase in MIC above that level is associated with unfavorable treatment outcomes (26). But, for

G en era l di sc us sio n a nd f ut ur e p er sp ec tiv

es drugs like isoniazid, rifampicin and FQs; strains with slightly elevated

MICs could still be treated by dose increment (26). This brings us to an important question “Can we afford to replace core drugs with less effective ones, if they can still be used in higher doses?”. The answer is clearly no. The PK/PD and/or clinical outcome data suggests that strains with a modest MIC increase above the critical concentration, might be treated with higher doses (26) which necessitates defining a clinical break point that distinguishes Mtb strains that are likely to succeed the therapy from those likely to fail (S-I-R system) (53). The phenotypic wild type strains are distinguished from non-wild type strains, with a possibility of treating the latter (modest MIC increase) using higher doses (26). The MIC for susceptible is up to 1 mg/L (range, 0.12–1) for Lfx in both MGIT/7H9 and 7H10 medium (26). For strains with MIC >0.5 mg/L, doses can be optimized based on measurement of plasma drug concentrations and individual MICs (Chapter 2). In line with our proposition, a recently published hollow fiber study found levofloxacin susceptibility breakpoint at an MIC of 0.5 mg/L (3). For strains with MIC of 1 mg/L, levofloxacin doses needed to be doubled to attain the optimal target exposure. This was also evident in a prospective study Figure 1: Flowchart for TDM of levofloxacin in programmatic settings.

7

MDR-TB patients with mycobacteria with higher MICs. Specific mutations at A90V, S91P, D94A confer to a low level of resistance with corresponding MIC values between 2–4 mg/L (MGIT/7H9 and 7H10 medium), in that case, levofloxacin should be better avoided and replaced by alternative drugs (26). The high-level resistance is characterized by mutations at D94G, which corresponds to an MIC above 8 mg/L. This level of mutations cannot not be compensated by increased dosing (26). The S-I-R system that is based on the phenotypic MIC prediction from genotypic mutations, requires as a foundation a good genotypic-phe-notypic correlation (15, 54).

At last, based on a review of available evidence and the results from clinical studies, the answers to the key clinical questions raised are discussed in the following section.

1. Is there a role of pharmacokinetic/pharmacodynamic (PK/PD) approaches in treatment optimization of Lfx? What should be the PK/PD target?

Yes, there is a direct relationship between Lfx sub-optimal concentra-tions, MIC of infecting Mtb and acquisition of drug resistance. This is also true for other anti-TB drugs such as moxifloxacin, isoniazid and rifampicin. Achieving optimal target exposure was positively associated with treatment success in MDR-TB patients based on observational studies. In a hollow fiber model of TB, AUC0–24/MIC >146 was asso-ciated with maximum Mtb kill and >360 prevented the development of acquired drug resistance.

2. Based on that target, can we use TDM to find the right dose?

Yes, TDM has a crucial role in addressing the effect of inter-individual pharmacokinetic variabilities in Cmax, AUC and Cmin in different

in-dividuals. Patients on the same standard dosage or mg/kg dosage may achieve a range of concentrations, which is difficult to predict without actually measuring it. Therefore, TDM helps to ensure adequate drug exposure.

Combining exposure data with a PD parameter then helps to deter-mine the optimal dose for the individual patient needed to kill the bug and prevent the development of acquired drug resistance.

G en era l di sc us sio n a nd f ut ur e p er sp ec tiv es

3. How feasible is TDM in resource limited settings using

conventional venous sampling? Can TDM be accomplished using alternative sampling strategies such as saliva?

TDM using venous sampling at all levels of care in TB endemic settings does not seem feasible at present due to its technical requirements. However, with the help of alternative and optimal sampling strategies, TDM could actually be implemented in the same manner TB diagnos-tics are. For Lfx, saliva sampling might not be a suitable matrix to accu-rately predict plasma concentrations for TDM due to a high variability in individual saliva-to-plasma ratios. Nevertheless, Lfx concentrations in saliva can be a valuable tool to pre-select patient’s sub-population eligible for TDM using a venapunction.

4. What is the optimal daily dose and dosing frequency of Lfx under programmatic conditions, to serve the majority of patients?

The bactericidal activity of Lfx is dependent on its concentration at the target site and its efficacy is predicted by AUC0–24/MIC ratio. Re-garding dosing frequency, once daily dosing should be preferred over administering the same dose in a divided fashion since AUC0–24 might be similar but Cmax is lower when total daily dose is divided. Based

on drug exposure data from available clinical studies and MIC distri-bution of clinical isolates, the currently used once daily 750–1000 mg dosing seems sub-optimal in the majority of patients and needs to be increased. The best tolerable Lfx dose with maximum efficacy for MDR-TB patients will be determined when results from the OptiQ trial are made available.

In future, adequately powered prospective clinical studies

vali-dating PK/PD targets from preclinical models (hollow-fiber and

animal models) to those observed in patients in clinics for

individ-ual TB drugs, and dose adjustments thereafter based on attainment

of established target could be a real shift towards personalized

medicine. TDM plays an important role in the selection of safe

and effective regimen for individual patients. Therefore, to make

TDM an attainable goal in remote settings, development of simple,

affordable, point-of-care test for determining drug concentrations

in alternative sampling matrices such as saliva should be a priority.

7

REFERENCES

(1) Zuur MA, Bolhuis MS, Anthony R, den Hertog A, van der Laan T, Wilffert B, et al. Current status and opportunities for therapeutic drug monitor-ing in the treatment of tuberculosis. Expert Opin Drug Metab Toxicol 2016;12(5):509–521.

(2) Dheda K, Gumbo T, Maartens G, Dooley KE, McNerney R, Murray M, et al. The epidemiology, pathogenesis, transmission, diagnosis, and manage-ment of multidrug-resistant, extensively drug-resistant, and incurable tuberculosis. Lancet Respir Med 2017;5(4):291–360.

(3) Deshpande D, Pasipanodya JG, Mpagama SG, Bendet P, Srivastava S, Koeuth T, et al. Levofloxacin pharmacokinetics/pharmacodynamics, dosing, sus-ceptibility breakpoints, and artificial intelligence in the treatment of mul-tidrug-resistant tuberculosis. Clin Infect Dis 2018;67(suppl_3):S293-S302. (4) Schentag JJ, Meagher AK, Forrest A. Fluoroquinolone AUIC Break Points and the Link to Bacterial Killing Rates: Part 1: In Vitro and Animal Models. Ann Pharmacother 2003;37(9):1287–1298.

(5) Schentag JJ, Meagher AK, Forrest A. Fluoroquinolone AUIC break points and the link to bacterial killing rates part 2: human trials. Ann Pharmacother 2003;37(10):1478–1488.

(6) Peloquin CA, Hadad DJ, Molino LP, Palaci M, Boom WH, Dietze R, et al. Pop-ulation pharmacokinetics of levofloxacin, gatifloxacin, and moxifloxacin in adults with pulmonary tuberculosis. Antimicrob Agents Chemother 2008 Mar;52(3):852–857.

(7) Thwaites GE, Bhavnani SM, Chau TT, Hammel JP, Torok ME, Van Wart SA, et al. Randomized pharmacokinetic and pharmacodynamic compari-son of fluoroquinolones for tuberculous meningitis. Antimicrob Agents Chemother 2011 Jul;55(7):3244–3253.

(8) van’t Boveneind-Vrubleuskaya N, Seuruk T, van Hateren K, van der Laan T, Kosterink JG, van der Werf, Tjip S, et al. Pharmacokinetics of Levoflox-acin in Multidrug-and Extensively Drug-Resistant Tuberculosis patients. Antimicrob Agents Chemother 2017:AAC. 00343–17.

(9) Nuermberger E, Grosset J. Pharmacokinetic and pharmacodynamic issues in the treatment of mycobacterial infections. Eur J Clin Microbiol Infect Dis 2004;23(4):243–255.

(10) Ghimire S, Sturkenboom M, Akkerman O, Bolhuis M, Simbar N, Alffenaar JWC. Preliminary PK and PK/PD report on gatifloxacin, moxifloxacin and levofloxacin: Systematic review to establish or revise the current recommended dose for TB treatment. 2016 Manuscript.

(11) Cegielski JP, Dalton T, Yagui M, Wattanaamornkiet W, Volchenkov GV, Via LE, et al. Extensive drug resistance acquired during treatment of multi-drug-resistant tuberculosis. Clin Infect Dis 2014 Oct 15;59(8):1049–1063. (12) Cegielski JP, Kurbatova E, van der Walt M, Brand J, Ershova J, Tupasi T, et al. Multidrug-resistant tuberculosis treatment outcomes in relation to treatment and initial versus acquired second-line drug resistance. Clin Infect Dis 2015;62(4):418–430.

G en era l di sc us sio n a nd f ut ur e p er sp ec tiv es

(13) World Health Organization. Rapid Communication: Key changes to treat-ment of multi-drug and rifampicin- resistant tuberculosis (MDR/RR-TB). 2018; Available at: http://www.who.int/tb/publications/2018/WHO_ RapidCommunicationMDRTB.pdf?ua=1. Accessed September, 2018. (14) Ahmad N, Ahuja SD, Akkerman OW, Alffenaar JC, Anderson LF, et al.

Treatment correlates of successful outcomes in pulmonary multidrug-re-sistant tuberculosis: an individual patient data meta-analysis. The Lancet 2018;392(10150):821–834.

(15) Rigouts L, Coeck N, Gumusboga M, de Rijk W, Aung K, Hossain M, et al. Specific gyrA gene mutations predict poor treatment outcome in MDR-TB. J Antimicrob Chemother 2015;71(2):314–323.

(16) Ängeby K, Juréen P, Kahlmeter G, Hoffner SE, Schön T. Challenging a dog-ma: antimicrobial susceptibility testing breakpoints for Mycobacterium tuberculosis. World Health Organ 2012;90(9):693–698.

(17) Niward K, Angeby K, Chryssanthou E, Paues J, Bruchfeld J, Jureen P, et al. Susceptibility testing breakpoints for Mycobacterium tuberculosis categorize isolates with resistance mutations in gyrA as susceptible to fluoroquinolones: implications for MDR-TB treatment and the definition of XDR-TB. J Antimicrob Chemother 2016 Feb;71(2):333–338. (18) U.S. Food and Drug Administration. FDA updates warnings for

fluoro-quinole antibiotics on risks of mental health and low blood sugar adverse reactions. 2018; Available at: https://www.fda.gov/newsevents/newsroom/ pressannouncements/ucm612995.htm. Accessed December, 2018. (19) European Medicines Agency. Quinolone- and fluoro-quinolone containing

medicinal products. 2018; Available at: https://www.ema.europa.eu/en/ medicines/human/referrals/quinolone-fluoroquinolone-containing-me-dicinal-products. Accessed December 2018, 2018.

(20) Fish DN, Chow AT. The clinical pharmacokinetics of levofloxacin. Clin Pharmacokinet 1997;32(2):101–119.

(21) Bouton TC, Phillips PP, Mitnick CD, Peloquin CA, Eisenach K, Patien-tia RF, et al. An optimized background regimen design to evaluate the contribution of levofloxacin to multidrug-resistant tuberculosis treat-ment regimens: study protocol for a randomized controlled trial. Trials 2017;18(1):563.

(22) Peloquin CA, Phillips PP, Mitnick CD, Eisenach K, Patientia RF, Lecca L, et al. Increased Doses Lead to Higher Drug Exposures of Levofloxacin for the Treat-ment of Tuberculosis. Antimicrob Agents Chemother 2018:AAC. 00770–18. (23) Kim HJ, Seo KA, Kim HM, Jeong ES, Ghim JL, Lee SH, et al. Simple and accurate quantitative analysis of 20 anti-tuberculosis drugs in human plasma using liquid chromatography-electrospray ionization-tandem mass spectrometry. J Pharm Biomed Anal 2015 Jan;102:9–16.

(24) Kuhlin J, Sturkenboom MG, Ghimire S, Margineanu I, van den Elsen, Simone HJ, Simbar N, et al. Mass spectrometry for therapeutic drug monitoring of anti-tuberculosis drugs. Clin Mass Spectrom 2018.

(25) Gumbo T. New susceptibility breakpoints for first-line antituberculosis drugs based on antimicrobial pharmacokinetic/pharmacodynamic science and

7

2010 Apr;54(4):1484–1491.

(26) Lange C, Alghamdi WA, Al‐Shaer MH, Brighenti S, Diacon AH, DiNardo AR, et al. Perspectives for personalized therapy for patients with multi-drug‐resistant tuberculosis. J Intern Med 2018.

(27) van Crevel R, van de Vijver S, Moore DA. The global diabetes epidemic: what does it mean for infectious diseases in tropical countries? Lancet Diabetes Endocrinol 2017;5(6):457–468.

(28) Daskapan A, de Lange WC, Akkerman OW, Kosterink JG, van der Werf, Tjip S, Stienstra Y, et al. The role of therapeutic drug monitoring in indi-vidualised drug dosage and exposure measurement in tuberculosis and HIV co-infection. Eur Respir J 2015;45(2):569–571.

(29) Leung CC, Yew WW, Mok TY, Lau KS, Wong CF, Chau CH, et al. Effects of diabetes mellitus on the clinical presentation and treatment response in tuberculosis. Respirology 2017;22(6):1225–1232.

(30) Alffenaar JC, Gumbo T, Aarnoutse RE. Acquired drug resistance: we can do more than we think! Clin Infect Dis 2014;60(6):969–970.

(31) van der Burgt EP, Sturkenboom MG, Bolhuis MS, Akkerman OW, Kosterink JG, de Lange WC, et al. End TB with precision treatment! Eur Respir J 2016 Feb;47(2):680–682.

(32) Alffenaar JC, Tiberi S, Verbeeck RK, Heysell SK, Grobusch MP. Therapeutic drug monitoring in tuberculosis: practical application for physicians. Clin Infect Dis 2016:ciw677.

(33) Alsultan A, Peloquin CA. Therapeutic Drug Monitoring in the Treatment of Tuberculosis: An Update. Drugs 2014;74(8):839–854.

(34) Kempker RR, Barth AB, Vashakidze S, Nikolaishvili K, Sabulua I, Tukvadze N, et al. Cavitary penetration of levofloxacin among patients with multidrug-re-sistant tuberculosis. Antimicrob Agents Chemother 2015;59(6):3149–3155. (35) Peloquin C. Use of therapeutic drug monitoring in tuberculosis patients.

CHEST Journal 2004;126(6):1722–1724.

(36) Van Altena R, De Vries G, Haar C, de Lange W, Magis-Escurra C, van den Hof S, et al. Highly successful treatment outcome of multidrug-resis-tant tuberculosis in the Netherlands, 2000–2009. Int J Tuberc Lung Dis 2015;19(4):406–412.

(37) Forsman LD, Bruchfeld J, Alffenaar JC. Therapeutic drug monitoring to prevent acquired drug resistance of fluoroquinolones in the treatment of tuberculosis. Eur Respir J 2017 APR;49(4):1700173.

(38) Ghimire S, Bolhuis MS, Sturkenboom MG, Akkerman OW, de Lange WC, van der Werf TS, et al. Incorporating therapeutic drug monitoring into the World Health Organization hierarchy of tuberculosis diagnostics. Eur Respir J 2016 Jun;47(6):1867–1869.

(39) van den Elsen SHJ, Sturkenboom MGG, Van’t Boveneind-Vrubleuskaya N, Skrahina A, van der Werf TS, Heysell SK, et al. Population Pharmaco-kinetic Model and Limited Sampling Strategies for Personalized Dosing of Levofloxacin in Tuberculosis Patients. Antimicrob Agents Chemother 2018 Nov 26;62(12):10.1128/AAC.01092–18. Print 2018 Dec.

G en era l di sc us sio n a nd f ut ur e p er sp ec tiv es

(40) Diel R, Vandeputte J, de Vries G, Stillo J, Wanlin M, Nienhaus A. Costs of tuberculosis disease in the European Union: a systematic analysis and cost calculation. Eur Respir J J2014;43(2):554–565.

(41) Marks SM, Flood J, Seaworth B, Hirsch-Moverman Y, Armstrong L, Mase S, et al. Treatment practices, outcomes, and costs of multidrug-resistant and extensively drug-resistant tuberculosis, United States, 2005–2007. Emerg Infect Dis 2014 May;20(5):812–821.

(42) Touw DJ, Neef C, Thomson AH, Vinks AA. Cost-effectiveness of therapeutic drug monitoring: a systematic review. Ther Drug Monit 2005;27(1):10–17. (43) van den Elsen SHJ, Akkerman OW, Huisman JR, Touw DJ, van der Werf TS,

Bolhuis MS, et al. Lack of penetration of amikacin into saliva of tuberculo-sis patients. Eur Respir J 2018 Jan 11;51(1):10.1183/13993003.02024–2017. Print 2018 Jan.

(44) Idowu OR, Caddy B. A review of the use of saliva in the forensic detection of drugs and other chemicals. J Forensic Sci Soc 1982 Apr;22(2):123–135. (45) Ghimire S., Maharjan B., Jongedijk E.M., Kosterink J.G.W., Ghimire G.R.,

Touw D.J., et al. Evaluation of saliva as a potential alternative sampling ma-trix for therapeutic drug monitoring of levofloxacin in MDR-TB patients. Antimicrob Agents Chemother 2019 Feb; 63 (5): 02379–18

(46) What is the right dose of isoniazid? The 49th Union World Conference on Lung Health- The Union; 2018.

(47) What is the right dose of rifampicin? The 49th Union World Conference on Lung Health- The Union; 2018.

(48) What is the right dose of pyrazinamide? The 49th Union Conference on Lung Health- The Union; 2018.

(49) Ghimire S, Maharjan B, Jongedijk EM, Kosterink J, Ghimire G, Touw D, et al. Levofloxacin pharmacokinetics and pharmacodynamics and outcome in MDR-TB patients. Eur Respir J 2019 Jan; 10.1183/13993003.02107–2018 (50) Ghiculescu R. Therapeutic drug monitoring: which drugs, why, when and

how to do it. Australian Prescriber 2008;31(2):42–44.

(51) World Health Organization. Technical report on critical concentrations for TB drug suceptibility testing of medicines used in the treatment of drug-re-sistant TB. 2018; Available at: https://www.who.int/tb/publications/2018/ WHO_technical_report_concentrations_TB_drug_susceptibility/en/. Ac-cessed December, 2018.

(52) Kahlmeter G. The 2014 Garrod Lecture: EUCAST–are we heading towards international agreement? J Antimicrob Chemother 2015;70(9):2427–2439. (53) Mouton JW, Brown D, Apfalter P, Canton R, Giske C, Ivanova M, et al. The role of pharmacokinetics/pharmacodynamics in setting clinical MIC break-points: the EUCAST approach. Clin Microbiol Infect 2012;18(3):E37-E45. (54) Coeck N, de Jong BC, Diels M, de Rijk P, Ardizzoni E, Van Deun A, et al.

Correlation of different phenotypic drug susceptibility testing methods for four fluoroquinolones in Mycobacterium tuberculosis. J Antimicrob Chemother 2016 May;71(5):1233–1240.