Pharmacokinetics of Ofloxacin and Levofloxacin for Prevention and Treatment of Multidrug-Resistant Tuberculosis in Children

Abstract

Limited data on fluoroquinolone pharmacokinetics and cardiac effects in children exist. Among 22 children receiving drug-resistant tuberculosis prophylaxis or treatment, serum concentrations following oral doses of levofloxacin (15 mg/kg of body weight) and ofloxacin (20 mg/kg) were lower than those expected from existing pediatric data, possibly due to differences in the formulations (crushed tablets). Drug exposures were lower than those in adults following standard doses and below the proposed pharmacodynamic targets, likely due to more rapid elimination in children. No QT prolongation was observed.

TEXT

Multidrug-resistant tuberculosis (MDR-TB) (i.e., resistance to rifampin and isoniazid) is an emerging epidemic with an estimated 63,000 pediatric cases per year (1, 2). Fluoroquinolones are essential in drug-resistant TB (DR-TB) treatment, but pharmacokinetic and safety data (including data on potential QT interval prolongation) for children are limited (3–6). South African DR-TB treatment guidelines were revised during 2012 to recommend levofloxacin (Lfx) and moxifloxacin (Mfx) instead of ofloxacin (Ofx) in children. Due to tablet sizes, children >8 years receive Mfx and children <8 years receive Lfx for prevention or treatment of DR-TB. We aimed to characterize the pharmacokinetics and cardiac effects of Ofx and Lfx in children <8 years of age.

A prospective crossover intensive pharmacokinetic sampling study was conducted at Brooklyn Chest Hospital and Tygerberg Children's Hospital in Cape Town, South Africa from May 2012 through March 2013. Children aged 3 months to 8 years routinely started on either Ofx or Lfx for the prevention or treatment of MDR-TB were eligible. Exclusion criteria were a hemoglobin level of <8 g/dl or a weight of <5 kg. Child contacts (<5 years or HIV infected) of infectious MDR-TB cases without TB disease received preventive therapy, including 20 mg/kg of body weight Ofx or 15 mg/kg Lfx daily for 6 months. Children with MDR-TB received MDR-TB treatment based on World Health Organization recommendations (7). South African Medicines Control Council-approved tablets of Ofx (Tarivid, 200-mg tablets; Sanofi-Aventis, Midrand, South Africa) and Lfx (250-mg tablets; Cipla, Cape Town, South Africa) were used.

Two pharmacokinetic samplings were done, alternately for Ofx and Lfx, both at steady state. Exact doses of Ofx of 20 mg/kg and of Lfx of 15 mg/kg were used, and the tablets were broken accordingly and weighed to the nearest 0.1 mg. After an overnight fast, tablets were given whole or crushed and suspended in water, and the dose taken was observed. Blood samples were collected predose and at 1, 2, 4, 6, and 8 h after dosing in EDTA-containing tubes and centrifuged, and plasma was separated and frozen within 30 min at −80°C. Lfx concentrations were determined using a validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay developed in the Division of Clinical Pharmacology, University of Cape Town, South Africa. Plasma samples were extracted, and chromatographic separation was achieved on a Gemini-NX 5-μm C18 (50-mm by 2-mm) analytical column. An AB Sciex API 3000 mass spectrometer was operated at unit resolution in the multiple reaction monitoring mode. The calibration curve fitted a linear (weighted by 1/concentration) regression over the range of 0.0781 to 5.0 μg/ml. Ofx plasma concentrations were determined using an LC-MS/MS method validated over the concentration range of 0.0781 to 20.0 μg/ml (8).

Noncompartmental analysis was used to calculate pharmacokinetic measures. The area under the curve from 0 to 8 h (AUC_0-8) and the AUC from 0 to 24 h (AUC_0-24) were calculated according to the trapezoidal rule. Extrapolated Lfx and Ofx concentrations at 24 h were estimated assuming first-order elimination. The pharmacodynamic indices AUC_0-24/MIC and the maximum concentration of drug in serum (C_max)/MIC ratios were calculated using published MIC₉₀ estimates for Ofx of 1.0 and 2.0 μg/ml and for Lfx of 0.5 and 1.0 μg/ml (9–12). Stata 12.1 (StataCorp, College Station, TX) was used for analysis. A 12-lead electrocardiogram (ECG) was done at 3 h postdose on each pharmacokinetic assessment day, and corrected QT intervals (QTc) were calculated (according to Fridericia [13]). A QTc of >450 ms was considered prolonged.

Parental written informed consent was obtained. The study was approved by the Stellenbosch University Human Research Ethics Committee.

Twenty-three children (median age, 3.14 years; interquartile range, 1.3 to 4.0 years) were enrolled and completed both pharmacokinetic assessments (Table 1). One child was excluded because of incorrect dosing of Lfx. Seven Ofx and 10 Lfx concentration measures sampled at time zero were listed as below the limit of quantitation and were taken as 0; there were no missing concentration measurements for Ofx or Lfx. Summary pharmacokinetic parameters are described in Table 2, with pharmacokinetic measures stratified by the clinical characteristics shown in Tables 3 (Ofx) and 4 (Lfx). Table 5 shows the estimated pharmacodynamic measures which favored Lfx. The mean QTc was 361 ms (standard deviation [SD], 37 ms) for Ofx and 369 ms (SD, 22 ms) for Lfx. No child in either group had a QTc of >450 ms.

TABLE 1.

Demographic and clinical characteristics of children receiving ofloxacin and levofloxacin for prevention or treatment of DR-TB^a

Characteristic	MDR disease group (n = 12) (n [%])	MDR PTb group (n = 11) (n [%])
Age at enrollment
0–2 yr	4 (36.4)	5 (45.5)
2–6 yr	3 (27.3)	6 (54.6)
≥6 yr	4 (36.4)	0 (0.0)
Male	6 (54.6)	7 (63.6)
Previous TB episode or treatment	4 (36.4)	0 (0.0)
Certainty of TB diagnosis (n = 12)
Bacteriological confirmation	5 (45.5)
Probable TB	5 (45.5)
Suspected TB	1 (9.1)
TB disease type (n = 12)
PTBc only	7 (63.6)
EPTBd only	0 (0.0)
PTB and EPTB	4 (36.4)
Reason DR PT started (n = 11)
DR contact only		8 (72.7)
DR contact and positive TSTe		3 (27.3)
HIV infected at baseline	4 (36.4)f	0 (0.0)
Weight-for-age Z score < −2.0 (n = 23)	5 (45.5)	1 (9.1)
Height-for-age Z score < −2.0 (n = 22)	6 (54.6)	2 (18.2)
Weight-for-length Z score < −2.0 (n = 22)	1 (9.1)	0 (0.0)

^aDR-TB, drug-resistant tuberculosis.

^bPT, preventive therapy.

^cPTB, pulmonary TB.

^dEPTB, extrapulmonary TB.

^eTST, tuberculin skin test.

^fNote that all HIV-infected children were >6 years of age.

TABLE 2.

Summary statistics for pharmacokinetic parameters of ofloxacin and levofloxacin in children receiving treatment or prophylaxis for multidrug-resistant tuberculosis (n = 22)^a

Drug	Cmax (μg/ml)	Tmax (h)	kel (1/h)	t1/2b (h)	AUC0–8 (μg · h/ml)	AUC(0–∞) (μg · h/ml)
Ofloxacin	9.75 (7.51–10.90)	1.59 (0.73)	0.22 (0.20–0.25)c	3.18 (2.83–3.44)d	43.84 (36.73–54.46)	47.51 (40.26–58.50)
Levofloxacin	6.79 (4.69–8.06)	1.45 (0.51)	0.22 (0.20–0.26)	3.16 (2.68–3.51)	30.47 (24.41–36.39)	32.92 (25.44–40.88)

^aAll parameters are reported using medians (interquartile ranges [IQRs]), except for time to maximum concentration in serum (T_max), which is reported using means (SDs).

^bt_1/2, half-life.

^cOne child was excluded from the analysis because a wrong dose of Lfx (10 mg/kg) was given.

^dOne child had a late peak in the concentration-time curve. For this reason, no half-life or elimination rate constant is reported for that child.

TABLE 3.

Pharmacokinetic measures of ofloxacin by study group, age, HIV status, and nutritional status (n = 22)

Clinical characteristic	Pharmacokinetic measure of ofloxacin
	Cmax			Tmax			AUC0–8			t1/2a
	No.	Median (IQR) (μg/ml)	P	No.	Mean (SD) (h)	P	No.	Median (IQR) (μg · h/ml)	P	No.	Median (IQR) (h)	P
Study group
MDR disease	11	10.20 (7.51–10.90)		11	1.82 (0.87)		11	46.53 (38.88–54.98)		10	3.13 (2.89–3.44)
MDR PTb	11	8.88 (7.05–12.70)	0.77c	11	1.36 (0.51)	0.15c	11	43.34 (29.75–50.89)	0.53c	11	3.18 (2.64–3.61)	0.89c
Age group
0–2 yr	9	10.90 (10.20–12.70)		9	1.22 (0.44)		9	46.53 (43.05–54.46)		9	2.91 (2.61–3.26)
2–6 yr	9	8.78 (5.39–9.82)		9	1.56 (0.53)		9	44.34 (28.99–48.76)		9	3.21 (3.00–3.61)
≥6 yrd	4	7.69 (6.21–9.39)	0.02e	4	2.5 (1.0)	0.01e	4	39.01 (33.47–48.06)	0.54c	3	3.44 (2.89–3.57)	0.29e
HIV status
HIV infectedf	4	7.69 (6.21–9.39)		4	2.5 (1.0)		4	39.01 (33.47–48.06)		3	3.44 (2.89–3.57)
HIV uninfected	18	9.86 (8.09–11.57)	0.25c	18	1.39 (0.50)	0.11c	18	44.67 (36.73–54.46)	0.55e	18	3.16 (2.63–3.34)	0.48c
WAZg
≥−2.0	18	10.05 (8.78–11.57)		18	1.44 (0.51)		18	45.77 (39.12–54.98)		18	3.20 (2.89–3.44)
<−2.0	4	6.63 (5.15–7.98)	0.04c	4	2.25 (1.26)	0.29c	4	29.37 (28.52–34.45)	0.03c	4	2.64 (2.56–3.57)	0.42c

^at_1/2, half-life. For one child, no t_1/2 could be reported due to a late peak serum concentration of Ofx.

^bPT, preventive therapy.

^cP value was calculated using the Mann-Whitney U test.

^dOne child was excluded from the analysis because a wrong dose of Lfx (10 mg/kg) was given.

^eP value was calculated using the Kruskal-Wallis test.

^fNote that all of the HIV-infected children were also >6 years old.

^gWAZ, weight-for-age Z score using WHO reference standards (29).

TABLE 4.

Pharmacokinetic measures of levofloxacin by study group, age, HIV status, and nutritional status (n = 22)

Clinical characteristic	Pharmacokinetic measure of levofloxacin
	Cmax			Tmax			AUC0–8			t1/2a
	No.	Median (IQR) (μg/ml)	P	No.	Mean (SD) (h)	P	No.	Median (IQR) (μg · h/ml)	P	No.	Median (IQR) (h)	P
Study group
MDR disease	11b	7.00 (4.69–8.06)		11b	1.46 (0.52)		11b	32.50 (24.41–38.83)		11b	3.24 (3.01–3.99)
MDR PTc	11	6.32 (4.63–8.17)	0.53d	11	1.46 (0.52)	1.00d	11	29.89 (21.07–32.75)	0.34d	11	2.98 (2.64–3.51)	0.31d
Age group
0–2 yr	9	7.00 (6.32–8.06)		9	1.33 (0.50)		9	29.89 (24.50–36.39)		9	2.79 (2.62–3.14)
2–6 yr	9	6.86 (4.69–7.51)		9	1.56 (0.53)		9	31.69 (23.81–33.11)		9	3.22 (2.98–4.24)
≥6 yre	4	4.98 (4.52–7.48)	0.40f	4	1.50 (0.58)	0.66f	4	27.49 (24.73–34.03)	0.91f	4	3.37 (3.12–4.01)	0.18f
HIV status
HIV-infectede	4	4.98 (4.52–7.48)		4	1.5 (0.58)		4	27.49 (24.73–34.03)		4	3.37 (3.12–4.01)
Not HIV-infected	18	6.88 (5.36–8.06)	0.39c	18	1.44 (0.51)	0.85c	18	31.38 (24.41–36.39)	0.67c	18	3.09 (2.64–3.51)	0.23c
WAZg
≥−2.0	17	6.86 (4.69–8.06)		17	1.47 (0.52)		17	31.06 (23.81–36.39)		17	3.14 (2.78–3.51)
<−2.0	5	6.71 (5.38–7.12)	0.91c	5	1.40 (0.55)	0.79c	5	29.24 (25.75–32.50)	0.91c	5	3.23 (2.68–3.51)	0.97c

^at_1/2, half-life.

^bOne child was excluded from the analysis because a wrong dose of Lfx (10 mg/kg) was given.

^cPT, preventive therapy.

^dP value was calculated using the Mann-Whitney U test.

^eNote that all of the HIV-infected children were also >6 years old.

^fP value was calculated using the Kruskal-Wallis test.

^gWAZ, weight-for-age Z score using WHO reference standards (29).

TABLE 5.

Estimated pharmacodynamic parameters using published MICs of ofloxacin and levofloxacin for Mycobacterium tuberculosis in children

Pharmacokinetic parameter	Estimate for 20 mg/kg ofloxacin with MIC (μg/ml) of:		Estimate for 15 mg/kg levofloxacin with MIC (μg/ml) of:		Pa
	2.0	1.0	1.0	0.5
AUC0–24/MIC (n = 21) (mean [SD])	31.5 (10.3)	63.0 (20.5)	43.8 (13.3)	87.6 (26.7)	<0.01
Cmax/MIC (n = 22) (mean [SD])	4.5 (1.5)	9.6 (3.1)	6.5 (2.0)	13.1 (4.0)	<0.01

^aPaired t test comparing pharmacodynamic parameters of Ofx and Lfx with an MIC of 2.0 μg/ml for Ofx versus an MIC of 1.0 μg/ml for Lfx and an MIC of 1.0 mg/ml for Ofx versus an MIC of 0.5 μg/ml for Lfx.

This study documents lower drug exposure to Ofx and Lfx in children compared to the expected values based on existing pediatric studies (4–6). This may be related to differences in drug formulations or administration methods. Tablets are the only oral formulation available in our setting; these need to be crushed for administration to young children. Although crushing might influence absorption, which is almost complete for the fluoroquinolones (4, 5, 14, 15), a study in adults comparing the pharmacokinetics of tablets and oral solutions of Ofx showed no differences (16). Although elimination tended to be higher in younger children, there were no differences for the Ofx and Lfx AUCs by age group. This may be related to confounding by HIV, as HIV infection is associated with reduced absorption of some anti-TB agents (17–19), and all the HIV-infected children in our study were >6 years of age. A larger study sample would better characterize the pharmacokinetics of these drugs and provide insights into any differences in this population compared to those in previous reports.

We also documented lower Ofx and Lfx exposures in children compared to those in adults using the currently recommended doses (20–22). Children eliminated Ofx and Lfx more rapidly than adults, with a half-life of 3 to 4 h and an elimination rate constant (k_el) of 0.22 h⁻¹ in our study compared to half-lives of 4 to 8 h and k_el of 0.09 to 0.13 h⁻¹ in adults (14, 20–22); this is likely due to the well-described age-related changes in drug clearance in children compared to drug clearance in adults (23). The proposed pharmacodynamic targets for fluoroquinolones against Mycobacterium tuberculosis are an AUC_0-24/MIC of >100 and C_max/MIC of 8 to 10 (12, 24–26). Children in this study failed to achieve an AUC_0-24/MIC of >100 even if a lower MIC₉₀ was used. Nevertheless, our estimated pharmacodynamic indices favor Lfx over Ofx. The failure to approximate adult exposures or to achieve the pharmacodynamic targets has important implications for current and future dosing recommendations in children for these essential anti-TB drugs. With the current recommended dosing, successful MDR-TB treatment outcomes are reported in 80 to 90% of children (27, 28). However, to optimize treatment, higher or more frequent dosing of Ofx and Lfx in young children may be required to approximate adult exposures and meet the proposed pharmacodynamic targets.

Although the ECG timing may have been suboptimal in our study and additional research is needed, the lack of any QT prolongation is reassuring.

The limitations of this study are the modest sample size, the unavailability of individual MICs, and the administration of different oral formulations of Ofx and Lfx. Additional studies on the safety and pharmacokinetics of Ofx and Lfx in children using higher doses to determine any impacts on treatment responses and outcomes should be considered.