INT J TUBERC LUNG DIS 23(10):1068–1074 Q 2019 The Union http://dx.doi.org/10.5588/ijtld.18.0775
Effect of tablet crushing on drug exposure in the treatment of multidrug-resistant tuberculosis
R. Court,1 M. T. Chirehwa,1 L. Wiesner,1 N. de Vries,2 J. Harding,3 T. Gumbo,4 G. Maartens,1 H. McIlleron1 1Division of Clinical Pharmacology, Department of Medicine, University of Cape Town, Cape Town, 2Brooklyn Chest Hospital, Cape Town, 3DP Marais Hospital, Cape Town, South Africa; 4Center for Infectious Diseases Research and Experimental Therapeutics, Baylor Research Institute, Baylor University Medical Center, Dallas, TX, USA
SUMMARY
SETTING: Treatment outcomes in multidrug-resistant compartmental analysis was used to derive the key tuberculosis (MDR-TB) are poor. Due to drug toxicity pharmacokinetic measurements. and a long treatment duration, approximately half of RESULTS: Twenty participants completed the study: 15 patients are treated successfully. Medication is often were men, and the median age was 31.5 years. There was crushed for patients who have difficulty swallowing a 42% reduction in the area under the curve AUC0–10 of whole tablets. Whether crushing tablets affects drug INH when the tablets were crushed compared with exposure in MDR-TB treatment is not known. whole tablets (geometric mean ratio 58%; 90%CI 47– OBJECTIVE AND DESIGN: We performed a sequential 73). Crushing tablets of pyrazinamide, moxifloxacin, pharmacokinetic study in patients aged .18 years on ethambutol and terizidone did not affect the bioavail- MDR-TB treatment at two hospitals in Cape Town, ability significantly. South Africa. We compared the bioavailability of CONCLUSION: We recommend that crushing of INH pyrazinamide, moxifloxacin, isoniazid (INH), ethambu- tablets in the MDR-TB treatment regimen be avoided. tol and terizidone when the tablets were crushed and Paediatric INH formulations may be a viable alternative mixed with water before administration vs. swallowed if the crushing of INH tablets is indicated. whole. We sampled blood at six time points over 10 h KEY WORDS: MDR-TB; crushed; pharmacokinetic; under each condition separated by 2 weeks. Non- bioequivalence; bioavailability
OUTCOMES FOR TREATMENT of multidrug- therefore, require drug administration via a nasogas- resistant tuberculosis (MDR-TB) are poor, with tric tube. treatment completion rates of ~54% being reported.1 However, tablet crushing may alter the bioavail- A heavy pill burden, together with nausea and ability of the active ingredients within the drug.7 vomiting, which has been reported to occur in Studies comparing the bioavailability of crushed vs. 75% of patients,2–4 contribute to poor regimen whole medication have shown that crushing decreas- tolerability.5 es the plasma concentrations of some drugs, including It is standard practice in some centres to crush rifapentine,8 but not of others.9–12 Combining the medication, and mix the crushed tablets with water to crushed tablets of a multidrug regimen is common; ease ingestion in the belief that this will reduce moreover, the tablets may be mixed into a vehicle- gastrointestinal upset. Crushing of tablets before containing substance that reacts with the drugs.12,13 administration is also common in young children Remnants may also adhere to the walls of the because suitable formulations are frequently not container in which the medication was crushed and available for those unable to swallow whole tablets. thereby escape ingestion. Subtherapeutic plasma One qualitative study at a paediatric hospital in Cape concentrations of some first and second-line anti- Town, South Africa, reported that 69% of caregiv- tuberculosis drugs have been associated with poor ers crush, dissolve or mix TB medication with food clinical outcomes, including acquisition of drug before administration.6 Crushing tablets may also be resistance.14–16 It is, therefore, important to under- necessary in critically ill patients with a depressed stand whether tablet crushing affects exposure of the level of consciousness who cannot swallow and, drugs used to treat MDR-TB, many of which are key
Correspondence to: Richard Court, Division of Clinical Pharmacology, University of Cape Town, Groote Schuur Hospital, K45 Old Main Building, Observatory 7925, South Africa. e-mail: [email protected] Article submitted 14 November 2018. Final version accepted 2 March 2019. Bioavailability of crushed anti-MDR-TB drugs 1069 drugs in the recently updated World Health Organi- water in a mixing cup. After ingestion, tablet zation (WHO) recommended management guide- remnants adhering to the walls of either the mortar, lines.17 pestle or mixing cup were scraped off with a spatula, mixed with a small unmeasured amount of water, and STUDY POPULATION AND METHODS swallowed by the participant. A standard breakfast was given to all participants 1 h after dosing. We We performed a sequential pharmacokinetic study. recorded all concurrent medication that could influ- This involved two intensive sessions of pharmacoki- ence plasma drug concentrations via drug-drug netic sampling in patients aged .18 years on MDR- interactions. At both pharmacokinetic sampling TB treatment at Brooklyn Chest Hospital and DP sessions, we sampled blood pre-dose as well as 2, 4, Marais Hospital in Cape Town. 6, 8 and 10 h post-dose. After centrifugation, plasma Between May 2016 and February 2017, we was extracted using a pipette and stored temporarily recruited participants with rifampicin (RMP) resis- on dry ice before being transported to the Division of tant TB who qualified for MDR-TB treatment. At the Clinical Pharmacology at the University of Cape time of the study, the standard MDR treatment Town for storage at 808C. Plasma drug concentra- regimen comprised pyrazinamide (PZA), moxiflox- tions were determined using liquid chromatography acin (MFX), kanamycin, cycloserine (CS) (dosed as tandem-mass spectrometry (LC-MS/MS).20–22 LC- terizidone [TRD]), and either ethionamide or isoni- MS/MS was validated according to guidelines set by azid (INH) depending on the results of the line-probe the US Food and Drug Administration and European assay for katG and inhA mutations identified in the Medicines Agency.23,24 pretreatment sputum culture, which indicated high- We used Stata v15.0 (Stata Corp, College Station, level resistance to INH or low-level resistance to INH TX, USA) to perform non-compartmental pharma- and resistance to ethionamide, respectively.18 Etham- cokinetic and statistical analyses. We determined the butol (EMB) was added if there had been no EMB following pharmacokinetic parameters for each drug exposure in the month before treatment initiation, on both dosing sessions: area under the concentration and the possibility of EMB resistance was considered time curve at 0–10 h (AUC ) using the trapezoidal to be low. We considered patients eligible for 0–10 rule, area under the concentration-time curve extrap- recruitment who were taking whole-tablet MDR olated to infinity (AUC ), half-life, peak concentra- treatment, either for MDR-TB (defined as resistance ‘ tion (C ) and time to C . to both RMP and INH)17 or for RMP-monoresistant max max We regarded pre-dose drug concentrations below TB (resistance to RMP but susceptible to INH). the lower level of quantification (BLQ) to be zero if Two sessions of pharmacokinetic sampling spaced all the pre-dose concentrations for a particular drug approximately 1–3 weeks apart (to allow as little inter-session variability as possible) were completed were BLQ. If any pre-dose concentrations for a drug for each participant 2 weeks after treatment were quantifiable, we then regarded all pre-dose initiation. Drug doses were in accordance with the BLQ concentrations for that drug to be half the national treatment guidelines during the study period lower level of quantification (LLQ). Similarly, post- and adjusted for toxicity before the first session of dose drug concentrations within the sampling pharmacokinetic sampling at the discretion of the interval were unlikely to be zero, so we considered treating clinician.19 any post-dose BLQ results for any of the drugs to be In the case of low-level INH resistance, participants half the LLQ. were given high doses of INH (10–15 mg/kg); the We used the Wilcoxon signed-rank test for paired standard dose of INH (5 mg/kg) was prescribed for data to compare Cmax and AUC0–10 at each session. participants with RMP-monoresistant TB. Partici- Then, the log-transformed values of Cmax and pants were given the same drug doses on both sessions AUC0–10 for exposure to crushed and whole tablets of pharmacokinetic sampling: whole tablets at the were compared with Student’s t-tests. The geometric first session and crushed tablets at the second session. mean ratio (GMR) point estimates and 90% Participants continued with whole-tablet treatment confidence intervals (CIs) of Cmax and AUC0–10 for until the second session of pharmacokinetic sampling, crushed vs. whole tablets were calculated for PZA, when they received crushed tablets; thereafter, they MFX, EMB, INH and CS. continued on whole tablet treatment. Dosing was Approval of our study protocol was granted by the performed under fasting conditions and was strictly Human Research Ethics Committee of the University observed by the study physician or nurse. of Cape Town, Cape Town, South Africa (106/2016). At the second session of pharmacokinetic sam- Written informed consent was taken from each pling, tablets were crushed with a standard-size participant in a language of his/her choice (English, mortar and pestle; TRD capsules were opened Afrikaans or Xhosa). All informed consent was carefully. All contents were mixed with 200 ml of obtained before participant recruitment. 1070 The International Journal of Tuberculosis and Lung Disease
Table 1 Characteristics of 20 patients on treatment for MDR-TB in a sequential comparative pharmacokinetic analysis
Participant characteristics Median [IQR] Men/women n/N 15/5 MDR-TB/rifampicin-monoresistant TB n/N 13/7 Age, years 31.5 [25.8–44.0] HIV-positive/negative n/N 10/10 Session 1 Session 2 Weight, kg 49 [44–54] 50 [44–55] BMI, kg/m2 17.4 [16.0–19.3] 17.1 [16.6–19.6] Creatinine clearance, ml/min* 94.4 [81.1–105.5] 92.1 [80.7–103.8] Duration on treatment at time of pharmacokinetic sampling, days 40.5 [32–45] 53 [44.5–60] Dose, mg/kg Pyrazinamide (n ¼ 20) 29.8 [27.8–30.9] 29.4 [27.3–31.1] Isoniazid (n ¼ 17) 11.8 [6.5–12.2] 11.4 [6.4–12.1] Moxifloxacin (n ¼ 20) 8.2 [7.6–9.0] 8 [7.4–9.1] Ethambutol (n ¼ 19) 16.7 [16–20] 17.0 [15.7–19.0] Terizidone (n ¼ 20) 15 [13.9–16.3] 15 [13.8–16.0]
* Cockcroft-Gault method. MDR-TB ¼ multidrug-resistant TB; IQR ¼ interquartile range; TB ¼ tuberculosis; HIV ¼ human immunodeficiency virus; BMI ¼ body mass index.
RESULTS and AUC‘ of some drugs could not be calculated in participants who had multiple drug concentration We recruited 25 participants, 20 of whom completed values reported as BLQ. Figure 1 shows the median the study: four completed only the first pharmacoki- time-concentration profiles of crushed and whole netic sampling session and one participant was tablets of INH to 10 h post-dose. The median time- withdrawn before any pharmacokinetic sampling concentration profiles of crushed and whole tablets of could be performed. CS, PZA, MFX and EMB are shown in Figure 2. The The characteristics of the 20 participants who increased variability of INH compared with the other completed the study are shown in Table 1. A drugs reflected the wider range of dosing, per INH descriptive comparison of the Cmax of each of the susceptibility, in participants with RMP-resistant drugs in crushed and whole form compared with TB.26 the expected range25 is shown in Table 2. Table 3 gives a comparison of the AUC0–10 and Cmax of whole and crushed tablets of PZA, MFX, EMB, DISCUSSION INH and CS. We noted significantly decreased exposure of INH The AUC0–10 of all drugs was reduced for crushed when the orally administered drugs in the MDR vs. whole-tablet formulations but the reduction was treatment regimen were crushed, and mixed with significant only for INH. We did not evaluate water. Dosing with crushed INH could affect ethionamide as too few participants (n ¼ 8) were outcomes in MDR-TB treatment, considering that receiving this drug at the time of pharmacokinetic low INH exposure has been associated with a poor sampling. The GMRs (with 90%CIs) for crushed vs. treatment response, including the development of whole tablets are shown in Table 4. Table 5 compares drug resistance.14–16 Exposure of the crushed-tablet additional pharmacokinetic parameters of the whole forms of the other drugs we assessed was also and crushed forms for each of the drugs. The half-life decreased compared with whole tablets at the same
Table 2 Comparison of Cmax of pyrazinamide, isoniazid, moxifloxacin, ethambutol and cycloserine with expected ranges in patients on treatment for multidrug-resistant tuberculosis
Whole-tablet Cmax Crushed-tablet Cmax (mg/l) (mg/l) Expected Cmax range Drug Median [IQR] Median [IQR] (mg/l) Pyrazinamide 42.7 [36.9–46.5] 4.0 [38.3–47.0] 20–60 Isoniazid Standard dose (5 mg/kg) (n ¼ 6) 0.9 [0.6–1.2] 0.6 [0.3–1.1] 3–6 High-dose (10–15 mg/kg) (n ¼ 11) 4.8 [3.5–6.9] 2.8 [2.1–4.2] Moxifloxacin (n ¼ 20) 2.4 [2.1–2.7] 2.3 [1.8–2.7] 3–5 Ethambutol (n ¼ 19) 1.9 [1.6–2.3] 1.8 [1.3–2.9] 2–6 Cycloserine (n ¼ 20) 33 [26.6–36.7] 34.5 [29.5–39.0] 20–35
Cmax ¼ peak concentration; IQR ¼ interquartile range. Bioavailability of crushed anti-MDR-TB drugs 1071
Table 3 Comparison of AUC0–10 and Cmax values between whole and crushed tablets in patients on treatment for multidrug- resistant tuberculosis
Whole-tablet Crushed-tablet Whole-tablet Crushed-tablet AUC0–10 AUC0–10 Cmax Cmax Drug median [IQR] median [IQR] P value median [IQR] median [IQR] P value Isoniazid (n ¼ 17) 13.8 [4.6–24.8] 7.3 [1.8–12.3] 0.02 3.5 [1.2–5.2] 2.1 [0.3–3.3] 0.02 Pyrazinamide (n ¼ 20) 316.1 [256.5–354.6] 307.0 [281.8–341.4] 0.35 42.7 [36.9–46.5] 41.0 [38.3–47.0] 0.13 Moxifloxacin (n ¼ 20) 15.2 [10.3–18.7] 14.2 [9.3–17.6] 0.22 2.4 [2.1–2.7] 2.3 [1.8–2.7] 0.06 Ethambutol (n ¼ 19) 11.3 [9.5–12.8] 11.0 [8.4–15.2] 0.63 1.9 [1.6–2.3] 1.8 [1.3–2.9] 0.75 Cycloserine (n ¼ 17) 281.9 [227.7–308.7] 281.2 [259.0–327.3] 0.49 32.7 [26.4–34.8] 34.3 [29.9–39.2] 0.39
AUC0–10 ¼ area under the concentration-time curve from 0 to 10 h; Cmax ¼ peak concentration; IQR ¼ interquartile range.
dose, but this difference did not reach statistical exposure of crushed tablets of INH being signifi- significance. cantly lower than the equivalent whole tablets. Our Little is known about the effect of crushing tablets finding of lower-than-expected exposures of whole on drug exposure in the treatment of MDR-TB. tablets of INH supports the notion of a possible Recently, a study in Cape Town observed low INH drug-drug interaction, which is enhanced if the exposures in children on MDR treatment, most of orally administered drugs are crushed together. We whom were dosed with crushed INH.27 INH is also a also considered that an interaction with an excipient key drug in the treatment of drug-susceptible and used in the production of one of the other drugs in drug-resistant TB, and is the drug of choice in TB the regimen, may be a possible cause of INH 28,29 chemoprophylaxis. These findings, therefore, degradation when the tablets were crushed togeth- have important clinical implications, particularly in er.32 Another potential reason, which could explain MDR-TB, where the companion drugs are relatively the reduced exposure of crushed tablets of INH, is less efficacious. that INH is considered by some scholars to be The mechanism causing poor INH exposure when unstable if mixed with water, although there are no MDR-TB drug formulations are crushed and mixed data to support this hypothesis. Conversely, crushed together in water is not known. TRD, which consists tablets of INH, if included in a fixed-drug combina- of two molecules of CS,30 has been reported to tion, have been shown to result in therapeutic interfere with INH absorption, but this effect is concentrations in the treatment of adults with poorly understood.31 A drug-drug interaction drug-susceptible TB.26 A powder formulation of whereby CS degrades INH before absorption or INH and crushed tablets of INH mixed with water inhibits INH absorption could be enhanced if the have also been shown to achieve target concentra- drugs are crushed and administered together in 33,34 water. We observed that exposure to whole tablets tions in children. of INH at standard doses to be lower than that In settings in which the crushing of INH tablets is reported in the literature (see Table 2), with indicated, there are several possible approaches to ensure optimal dosing. Paediatric INH formulations or constituting the dose with smaller INH tablets (e.g., 100 mg) with proven bioequivalence that are swallowed more readily35 should be used instead of crushing INH tablets in adults on treatment for MDR-TB. The extent to which the INH dose must be increased if crushed tablets of INH are adminis- tered to patients is unclear and requires further
study. We found a 42% reduction in the AUC0–10 of INH when the tablets were crushed compared with that using whole tablets (GMR 58%; 90%CI 47– 73%), indicating that the dose of INH, if crushed together with the other drugs in the MDR-TB treatment regimen, will require significant adjust- Figure 1 Concentration-time profile of crushed vs. whole ment to achieve target concentrations. Exposures to tablets of isoniazid in 17 participants on treatment for PZA, MFX, EMB, or CS were not significantly multidrug-resistant tuberculosis. Upper and lower bound of decreased by tablet crushing, indicating that tablets whiskers¼upper and lower interquartile range, respectively. The median concentrations of crushed and whole tablets at each of these drugs may be safely crushed when necessary. time point have been offset for clarity. Nevertheless, there was a trend towards lower 1072 The International Journal of Tuberculosis and Lung Disease
Table 4 Geometric mean ratios (90%CIs) of AUC0–10 and Cmax for crushed vs. whole tablets in the treatment of patients with multidrug-resistant tuberculosis
Isoniazid Moxifloxacin Pyrazinamide Ethambutol Cycloserine (n ¼ 17) (n ¼ 20) (n ¼ 20) (n ¼ 19) (n ¼ 20) % (90%CI) % (90%CI) % (90%CI) % (90%CI) % (90%CI)
AUC0–10 58 (47–73) 89 (80–99) 98 (93–103) 100 (89–112) 101 (91–113) Cmax 54 (40–73) 90 (82–98) 97 (93–101) 101 (85–121) 102 (91–114)
CI ¼ confidence interval; AUC0–10 ¼ area under the concentration-time curve to 10 h; Cmax ¼ peak concentration.
exposure of crushed MFX tablets compared with the effect on the AUC of the crushed tablets, as there equivalent whole tablets. was insufficient time for complete washout of drugs Our study had three main limitations. First, when with longer half-lives, particularly CS. tablets were crushed, care was taken to ensure that as much of the crushed tablet remnants as possible CONCLUSIONS were ingested by the participants by rinsing the mixing cup. We have observed that nursing staff in We recommend that crushing tablets of INH together busy treatment centres often do not have time to with the other orally administered drugs in the MDR ensure that all crushed remnants are swallowed by treatment regimen be avoided. Also, paediatric INH patients, which could result in a further reduction formulations should be considered for use in adults if in drug exposure. Our study may have been more tablet crushing is indicated. clinically relevant if we did not rinse the mortar, Acknowledgements pestle and mixing cup after crushing the tablets on the second day of pharmacokinetic sampling. This study was supported by a grant from the National Institute of Second, the pharmacokinetic sampling sessions Allergy and Infectious Diseases of the National Institutes of Health (NIH), Bethesda, MD, USA (R01AI116155 to HM and TG). The using crushed and whole tablets were not rando- University of Cape Town (UCT) Clinical Pharmacokinetic Labora- mised. We, therefore, cannot exclude the possibility tory is also supported by the National Institute of Allergy and of a sequence effect on our bioavailability compar- Infectious Diseases (NIAID) of the NIH under award numbers UM1 isons of crushed- and whole-tablet treatments. AI068634, UM1 AI068636 and UM1 AI106701. Overall support for the International Maternal Paediatric Adolescent AIDS Clinical Third, the pre-dose sample on the second crushed Trials Group (IMPAACT) at UCT was provided by the NIAID (U01 pharmacokinetic sample may have been affected by AI068632), The Eunice Kennedy Shriver National Institute of Child the pre-dose sample, which may have had some Health and Human Development, and National Institute of Mental
Figure 2 Concentration-time profiles of crushed vs. whole tablets of cycloserine, pyrazinamide, moxifloxacin and ethambutol in the treatment of patients* with multidrug-resistant tuberculosis. *n ¼ 20 unless otherwise indicated. Upper and lower bound of whiskers ¼ upper and lower interquartile range, respectively. The median concentrations of crushed and whole tablets at each time point have been offset for clarity. Cycloserine and pyrazinamide concentrations are 10-fold the concentration displayed on the y-axis. Bioavailability of crushed anti-MDR-TB drugs 1073
Health (AI068632) all of which are part of NIH. The content of this 17) article is solely the responsibility of the authors and does not ¼
n necessarily represent the official views of the NIH. HM and GM are also supported by the National Research Foundation of South Africa (90729 and 85810 respectively). HM is also supported by the median [IQR] Crushed-tablet Wellcome Trust (206379/Z/17/Z). We acknowledge the contribu-
,h tions of the patients who volunteered for our study. max
17) 2 [2–2] ( References
¼ 1 World Health Organization. Global tuberculosis report, 2017. n WHO/HTM/TB/2017.23. Geneva, Switzerland: WHO, 2017: p 248. http: //www.who.int/tb/publications/global_report/ Whole-tablet median [IQR] gtbr2017_annex4.pdf?ua¼1 2 Furin JJ, Mitnick CD, Shin SS, et al. Occurrence of serious adverse effects in patients receiving community-based therapy for multidrug-resistant tuberculosis. Int J Tuberc Lung Dis
13) 2 [2–2] ( 2001; 5(7): 648–655.
¼ 3 Isaakidis P, Varghese B, Mansoor H, et al. Adverse events n among HIV/MDR-TB co-infected patients receiving antiretroviral and second-line anti-TB treatment in Mumbai, India. PLoS One 2012; 7(7): e40781. 4 Shin SS, Pasechnikov AD, Gelmanova IY, et al. Adverse median [IQR] Crushed-tablet reactions among patients being treated for MDR-TB in Tomsk, Russia. Int J Tuberc Lung Dis 2007; 11(12): 1314–1320. 5 Isaakidis P, Rangan S, Pradhan A, Ladomirska J, Reid T, Kielmann K. ‘‘I cry every day’’: experiences of patients co- infected with HIV and multidrug-resistant tuberculosis. Trop
16) 2.5 [1.9–3.3] ( Med Int Health 2013; 18(9): 1128–1133.
¼ 6Belard´ S, Isaacs W, Black F, et al. Treatment of childhood n tuberculosis: caregivers’ practices and perceptions in Cape Town, South Africa. Paediatr Int Child Health 2015; 35(1): 24– 28.
Whole-tablet median [IQR] 7 Royal Pharmaceutical Society. Pharmaceutical issues when crushing, opening or splitting oral dosage forms. RPS, 2011. https://www.rpharms.com/Portals/0/RPS document library/ Open access/Support/toolkit/pharmaceuticalissuesdosage- forms-%282%29.pdf Accessed August 2019.
time to peak concentration. 8 Weiner M, Savic RM, Kenzie WRM, et al. Rifapentine ¼ 13) 2.4 [1.6–3.6] ( pharmacokinetics and tolerability in children and adults max ¼ treated once weekly with rifapentine and isoniazid for latent n tuberculosis infection. J Pediatric Infect Dis Soc 2014; 1–14. 9 Best B, Capparelli E, Diep H, et al. Pharmacokinetics of lopinavir/ritonavir crushed versus whole tablets in children. J Acquir Immune Defic Syndr 2012; 58(629): 385–391. median [IQR] Crushed-tablet 10 Argenti D, Ireland D, Heald DL. A pharmacokinetic and
interquartile range; T pharmacodynamic comparison of desmopressin administered ¼ h/ml Half-life, h T