Pharmacokinetics of Tedizolid in Morbidly Obese and Covariate-Matched Nonobese Adults

Abstract

Tedizolid is a novel oxazolidinone antimicrobial administered in its prodrug form, tedizolid phosphate, as a fixed once-daily dose. The pharmacokinetics of tedizolid has been studied in a relatively small proportion of morbidly obese (body mass index [BMI] of ≥40 kg/m²) adults through population analyses with sparse sampling. The current study compared the intensively sampled plasma pharmacokinetics of tedizolid phosphate and tedizolid in 9 morbidly obese and 9 age-, sex-, and ideal body weight-matched nonobese (BMI, 18.5 to 29.9 kg/m²) healthy adult (18 to 50 years of age) volunteers after administration of a single intravenous dose of tedizolid phosphate. The median (range) weights were 72.6 kg (58.9 to 89.5 kg) and 117 kg (102 to 176 kg) for the mostly female (77.8%) nonobese and morbidly obese adults, respectively. Tedizolid phosphate concentrations were below the limit of quantitation in a majority of subjects after the 2-h time point. The tedizolid median (range) maximum concentration of drug in plasma (C_max) and area under the concentration-time curve from 0 h to infinity (AUC_0–∞) were 2.38 (1.28 to 3.99) mg/liter and 26.3 (18.4 to 43.2) h · mg/liter, respectively, for morbidly obese subjects, and these were nonsignificantly different (P ≥ 0.214) from the values for nonobese subjects. Similarly, the volumes of distribution (V_z) (P = 0.110) and clearance (CL) values (P = 0.214) were comparable between groups. Nearly identical (P = 0.953) median tedizolid half-lives of approximately 12 h were observed for both groups. Tedizolid V_z and CL scaled with body weight, but not proportionately. The small and nonsignificant differences in tedizolid AUC_0–∞ values between morbidly obese and nonobese subjects suggest that dose modification is not necessary for morbidly obese adults. (This study has been registered at ClinicalTrials.gov under number NCT02342418.)

INTRODUCTION

More than one-third of the U.S. adult population is classified as obese, based on having a body mass index (BMI) of ≥30 kg/m² (1). Acute bacterial skin and skin structure infections (ABSSSI) are common in obese patients, who are predisposed to developing type 2 diabetes (2). Tedizolid is a novel once-daily oxazolidinone that was recently approved for the treatment of ABSSSI at 200 mg daily for 6 days. Tedizolid is a more potent oxazolidinone than linezolid, and it does not interact with selective serotonin reuptake inhibitors (SSRIs). This pharmacological difference is significant because the use of antidepressants is higher among obese adults than among nonobese adults in the United States (3, 4). As a consequence, tedizolid may be selected over linezolid for the treatment of ABSSSI in obese patients, who are potentially managed with SSRIs. To date, population pharmacokinetic (POP-PK) analyses have demonstrated that the concentration-time profiles of tedizolid are similar for subjects with class II obesity (BMI of ≥35 kg/m²) and nonobese adults (BMI of <30 kg/m²), suggesting that no dose modification for body size is necessary (5). In addition, POP-PK analyses clearly demonstrated that the tedizolid volume of distribution (V) and clearance (CL) scale as a power term function of ideal body weight (IBW) but not with body weight or BMI (6). POP-PK analysis suggested that identical doses of tedizolid can be administered to obese and nonobese adult patients of similar stature (6). However, the POP-PK model likely included a relatively small proportion of morbidly obese (BMI of ≥40 kg/m²) adults, as only one of two phase 3 trials enrolled patients (27 of 664 subjects) in this body size stratum (5–7).

Two phase 3 studies, TR701-112 and TR701-113, evaluated the efficacy, safety, and POP-PK of tedizolid at 200 mg once daily for 6 days compared to linezolid at 600 mg twice daily for 10 days (7). The primary efficacy outcome was an early clinical response, defined as a ≥20% reduction in lesion size, and apyrexia at 48 to 72 h. In TR701-112, patients with a BMI of 35 to 40 kg/m² had a lower clinical response rate with tedizolid than that with linezolid (77.9% versus 86.7%). Patients who were in the tedizolid group and had a BMI of >40 kg/m², who were recruited only in the TR701-113 trial, had an even lower clinical response rate (59.3% versus 80.8% for linezolid) (7). The 95% confidence interval for the response rate difference between these groups crossed zero (not significant), but a trend in the size of the difference was notable across body sizes (7). A subgroup analysis of the primary endpoint by medical history for the pooled studies showed that patients with diabetes who received tedizolid also had a numerically lower response rate than that of patients on linezolid. Although the studies found tedizolid to be noninferior to linezolid for the primary efficacy outcome, the numerical differences in the diabetic and obese populations cannot be ignored (7). The results of the analyses, albeit insufficiently powered, suggested the potential for different outcomes in obese patients, especially among those with diabetes.

The results of these subgroup analyses prompted a closer examination of the tedizolid dosing schedule in obese adults. The tedizolid areas under the concentration-time curves (AUC) and maximum concentrations of drug in plasma (C_max) were similar after single and multiple doses of intravenous tedizolid, which supports the role of a single-dose PK study for characterizing the concentration-time profile of this agent (8). If the tedizolid PK parameters are correlated to ideal body weight as suggested by POP-PK analyses, then similar exposures should be expected in morbidly obese and nonobese adults with the same dosage. Thus, the current study was designed to compare the single-dose (200 mg) plasma pharmacokinetics of intravenous tedizolid in morbidly obese (≥40 kg/m²) and age-, sex-, and ideal body weight-matched nonobese (18.5 to 29.9 kg/m²) adult volunteers. The correlations of tedizolid pharmacokinetic system parameters to body size parameters were also examined.

MATERIALS AND METHODS

Regulatory review.

The current study met the requirement for a waiver of an investigative new drug application (IND exemption no. 125012). The study was approved by the Albany College of Pharmacy and Health Sciences Institutional Review Board (IRB) and through IntegReview (Austin, TX). This clinical trial was performed at the clinical research unit of TKL Research (Rochelle Park, NJ) and was registered through ClinicalTrials.gov (ClinicalTrials registration number NCT02342418).

Inclusion criteria.

Subjects fulfilling the following criteria were eligible: (i) male or female and aged 18 to 50 years; (ii) nonsmoking or light smoking (≤5 cigarettes per day); (iii) estimated creatinine clearance (Cockcroft-Gault equation [9]) of ≥90 ml/min; (iv) female subjects of childbearing potential must have been surgically sterilized or using hormonal contraceptives or an effective barrier method of contraception (diaphragm, cervical cap, or condom) or must have agreed to abstain from sex from the time of prestudy screening through the entire study period and 4 weeks following the study period; (v) platelet count of ≥140,000/μl; and (vi) absolute neutrophil count of ≥1,800/μl.

Exclusion criteria.

Subjects fulfilling the following criteria were excluded: (i) history of hypersensitivity reaction to any oxazolidinone; (ii) BMI of <18.5 kg/m² or 30 to <40 kg/m²; (iii) any chronic medical condition requiring pharmacologic therapy; (iv) aspartate transaminase (AST) or alanine aminotransferase (ALT) level of >2.5 times the upper limit of normal; (v) total bilirubin level of >1.5 times the upper limit of normal; (vi) positive urine pregnancy test (if female); (vii) abnormal electrocardiogram as judged by the study physician; (viii) inability to tolerate venipuncture and multiple blood draws; (ix) clinically significant abnormal physical examination, defined as a physical finding requiring further clinical work-up; and (x) inability to independently provide written informed consent.

Sample size and subject-matching criteria.

Nine morbidly obese male and female subjects (BMI of ≥40 kg/m²) were recruited first. Each morbidly obese subject was matched to a nonobese subject (BMI of 18.5 to 29.9 kg/m²) based on age (±5 years), sex, and ideal body weight (±4.6 kg, i.e., ±5.08 cm in height) (10).

Pharmacokinetic sampling and bioanalysis.

A single 200-mg intravenous dose (1-h infusion) of tedizolid phosphate (Sivextro) was administered, with subsequent collection of 11 blood samples per subject over 72 h. The exact sampling time points included time zero (within 30 min of dosing) and 0.5, 1, 2, 4, 6, 8, 12, 24, 48, and 72 h after the start of infusion. Blood sampling was performed through a peripheral venous catheter in the arm contralateral to the site of drug infusion. The study subjects were discharged from the clinical research unit after the 12-h sample was collected, and they returned to the site for subsequent sampling by peripheral venous collection. All 5-ml blood samples were collected using K₃ EDTA blood collection tubes (BD, Franklin Lakes, NJ) to prevent ex vivo conversion of tedizolid phosphate to tedizolid. Each blood sample tube was inverted and turned upright 5 times to afford mixing of blood with the anticoagulant, maintained on wet ice, and centrifuged at 1,200 × g for 10 min at 4°C within 60 min of collection. Plasma samples were stored frozen at −70°C until bioanalysis. Sample analysis was performed by a validated method at Covance Laboratories (Madison, WI) as previously described (8). Briefly, tedizolid phosphate and tedizolid were assayed by liquid chromatography-tandem mass spectrometry, using labeled versions of both analytes as internal standards. The curve range of detection was 0.005 mg/liter to 1 mg/liter, and a maximum dilution factor of up to 50-fold was validated for both tedizolid and tedizolid phosphate. The intraday assay precision relative standard deviation was 2.2% to 9.1% across the low (0.015 mg/liter), medium (0.150 mg/liter), and high (0.750 mg/liter) quality control tedizolid concentrations, with a mean accuracy range of 99.2% to 107%. The interday assay precision relative standard deviation was 5.1% to 7.0% across the quality control tedizolid concentrations, with a mean accuracy range of 97.7% to 105%.

Pharmacokinetic and statistical analyses.

Noncompartmental pharmacokinetic analysis (NCA) was performed using Phoenix WinNonLin, version 6.4 (Mountain View, CA), given the data-rich nature of the sampling schema. The following parameters were derived by NCA: maximum concentration (C_max), time to maximum concentration (T_max), last measurable concentration (C_last), time to last measurable concentration (T_last), area under the curve from time zero to the last measurable concentration (AUC_0–last), area under the curve from time zero to infinity (AUC_0–∞), volume of distribution during the terminal phase after intravenous administration (V_z), clearance (CL), and terminal half-life. Between-group comparisons of matched subjects were performed using the Wilcoxon matched-pair signed-rank test and Fisher's exact test for categorical variables. Visual inspection of the scatterplot matrix, Pearson's correlation (R), and ordinary least-square regression were used to assess the relationships between pharmacokinetic system parameters (CL and V_z) and body size. In addition, nonlinear regression (power function) was used to compare pharmacokinetic system parameters to weight^β and α(weight/median weight)^β, where β is an exponent and α is a constant function of weight. Discrimination between models was performed based on the coefficient of determination (R²) and the Akaike information criterion (AIC). All statistical analyses were implemented through STATA SE, version 13.1.

RESULTS

Subject demographics.

A total of 19 subjects were enrolled in the study, with one withdrawal due to an intravascular infiltration during infusion of the dose. Demographic data based on body size categorical groups are summarized in Table 1. As expected based on the study inclusion criteria, the mean body weight and body mass index were approximately 1.62-fold higher in the morbidly obese group than in the nonobese group. In addition, the mean body surface area was 1.27-fold higher in the morbidly obese group than in the nonobese group.

TABLE 1.

Summary of population demographic parameters for nonobese and morbidly obese subjects

Parameter	Valuea		P value
	Nonobese subjects (n = 9)	Morbidly obese subjects (n = 9)
Age (yr)	37 (24–46)	38 (26–50)	0.120
Height (cm)	164 (144–182)	160 (154–185)	0.953
Weight (kg)	72.6 (58.9–89.5)	117 (102–176)	0.008
BMI (kg/m2)	27.4 (25.4–29.1)	43.5 (40.4–51.8)	0.008
BSA (m2)	1.82 (1.57–2.09)	2.26 (2.12–3.01)	0.008
IBW (kg)	56.4 (37.5–77.6)	52.4 (47.2–79.9)	0.953
Sex (no. of males/no. of females)	2/7	2/7	1.00
Race (no. of subjects)			1.00
White	5	5
Black	4	3
East Indian	0	1

^aData are medians (ranges) unless otherwise indicated.

Tedizolid phosphate concentration-time profile.

Tedizolid phosphate concentrations were measurable in all subjects at the midpoint and the end of the infusion but were measurable in only 3 nonobese subjects and 1 morbidly obese subject at the 2-h time point. The mean (standard deviation) tedizolid phosphate concentrations at the end of infusion were 1.73 (0.775) mg/liter and 0.909 (0.417) mg/liter (P = 0.0041) for the nonobese and morbidly obese groups, respectively. However, this potential difference was not relevant by the 2-h time point given that the majority of subjects had concentrations below the lower limit of quantification (LLOQ). The limited detection of tedizolid phosphate concentrations in plasma is consistent with the expectation of rapid conversion to tedizolid in plasma. Given the limited measurement profile of tedizolid phosphate, pharmacokinetic analyses were not performed.

Tedizolid concentration-time profile.

Tedizolid concentrations were measurable in all subjects and followed a monoexponential decline after the end of infusion. The mean and standard deviation concentration-time profile of tedizolid is illustrated for the nonobese and morbidly obese groups (Fig. 1). As shown in Fig. 1, mean concentrations were initially numerically higher in nonobese subjects than in morbidly obese subjects. However, no significant differences between the two groups were observed for the C_max, T_max, C_last, and T_last measurements (Table 2). The 200-mg dosage of tedizolid phosphate is equivalent to 164.5 mg of tedizolid, so this value was used to define the tedizolid dose (100% conversion was assumed). The median (range) values for the pharmacokinetic system parameters and measures of exposure are also included in Table 2. No statistically significant differences were observed between the groups. The rapid decline in tedizolid phosphate concentrations by 2 h corresponded to mean times to maximum concentration of 1.98 h and 1.42 h for nonobese and morbidly obese subjects, respectively (P = 0.214).

FIG 1.

Tedizolid plasma concentration-time profiles for nonobese and morbidly obese subjects. Data are means and standard deviations.

TABLE 2.

Tedizolid plasma pharmacokinetic parameters for nonobese and morbidly obese subjects

Parameter	Valuea		P value
	Nonobese subjects	Morbidly obese subjects
Cmax (mg/liter)	2.96 (2.16–5.27)	2.38 (1.28–3.99)	0.214
Tmax (h)	1.93 (0.917–3.98)	1.08 (0.917–2.00)	0.214
Clast (mg/liter)	0.017 (0.007– 0.055)	0.020 (0.014–0.057)	0.953
Tlast (h)	71.8 (48.8–73.0)	71.5 (47.8–72.8)	0.139
AUC0–last (h · mg/liter)	27.2 (22.4–56.0)	26.0 (18.0–42.0)	0.314
AUC0–∞ (h · mg/liter)	27.4 (22.7–57.0)	26.3 (18.4–43.2)	0.314
CL (liters/h)	5.99 (2.88–7.25)	6.25 (3.81–8.93)	0.214
Vz (liters)	88.2 (53.3–124)	101 (76.8–150)	0.110
Half-life (h)	11.7 (9.00–15.0)	11.9 (8.01–14.1)	0.953

^aData are medians (ranges).

Relationship of tedizolid pharmacokinetic parameters to body size.

The relationships of tedizolid CL to age, height, weight, ideal body weight (IBW), adjusted body weight (ABW), lean body weight (LBW), BMI, and body surface area (BSA) were assessed visually through scatterplots and linear correlation (10–13). Scatterplots of the relationships between CL, V_z, and weight are illustrated in Fig. 2. Based on a review of these scatterplots, a potential outlier (a 178-kg individual) may have served as a leverage point, so inclusion and exclusion of this subject were tested for impacts on the observed relationships. Inclusion and exclusion of this outlier had no impact on the intercept or slope of the relationships between these parameters (data not shown). Nonlinear regression of CL to weight^β revealed a mean (95% confidence interval) β coefficient of 0.396 (0.370, 0.422) (AIC = 66.3). The relationship of CL to α(weight/95.5)^β was also tested, but the AIC (68.2) was not improved. Regression of V_z to weight^β revealed a mean (95% confidence interval) β coefficient of 0.996 (0.971, 1.02) (AIC = 166). However, in this case, the relationship of tedizolid V_z to weight based on the function α(weight/95.5)^β was improved (AIC = 160). The mean (95% confidence interval) α coefficient was 101 (91.0, 110), and the mean (95% confidence interval) β coefficient was 0.527 (0.197, 0.858). Tedizolid CL had similar correlations with BSA (R = 0.530), ABW (R = 0.530), LBW (R = 0.521), and weight (R = 0.503). Similarly, tedizolid V_z also demonstrated good correlations with BSA (R = 0.684), ABW (R = 0.714), LBW (R = 0.753), and weight (R = 0.646).

FIG 2.

Scatter, linear, and power function regression plots with equations and coefficients of determination for tedizolid clearance and body weight (A) and volume of distribution and body weight (B).

Safety and tolerability.

Three subjects experienced four adverse events in total. One subject experienced an intravenous infiltration during the infusion. No action was required, as the infusion was terminated as soon as the adverse event was observed. Given the interruption in the infusion and the inability to define the dose received, the subject was withdrawn from the study after resolution of the adverse event. A subject experienced a headache that resolved with self-administration of ibuprofen. This subject also experienced six episodes of diarrhea that self-resolved by the end of the study visit. These two adverse events were judged by the study physician to probably be related to the study drug. Finally, one subject experienced an ankle injury secondary to a fall with no loss of consciousness. This event was self-managed by the subject and deemed unrelated to the study drug.

DISCUSSION

This is the first study to directly compare the pharmacokinetics of tedizolid in morbidly obese and nonobese adults. The study was performed to ascertain whether a dose adjustment of tedizolid phosphate is necessary in morbidly obese adults to achieve tedizolid exposures comparable to those in nonobese adults. The study was designed to match subjects by age, sex, and ideal body weight (i.e., height), because previous analyses suggested that ideal body weight was the optimal body size predictor of tedizolid pharmacokinetic system parameters. This study demonstrates that morbidly obese subjects matched by these criteria have nonsignificant differences in tedizolid exposure despite having a 62.2% higher mean body weight than that of nonobese subjects. Validation of this POP-PK expectation through this independent study and evaluation is important because it helps to ensure that a fixed dosing recommendation for tedizolid is reasonable across the wide adult patient weight range.

The plasma concentration-time profiles of tedizolid were similar for the nonobese and morbidly obese subjects. As expected with most drugs, the plasma C_max was numerically lower in morbidly obese subjects due to an expected increase in the volume of distribution. Analysis of the relationship of tedizolid V_z to body size clearly demonstrated a linear relationship (not proportionate) with body weight. Most importantly, tedizolid V_z was shown to scale linearly with body weight and had a β value of 1.0, which meets the allometric expectation. Previous analyses identified central compartment volume as a function of IBW, with a β value of 1.32. This finding is consistent with the results of the present study, because IBW is a function of height and sex and is related to weight as a supralinear function. Ultimately, less than 37% of interindividual variability of tedizolid V_z was explained by body weight. In simpler terms, 65-kg and 130-kg subjects would be expected to have V_z of 83.4 liters and 118 liters, respectively, which is a 26.3% difference despite a 2-fold difference in body weight. Given that the pharmacodynamic effects of tedizolid are not known to be correlated with C_max, the practical effects of weight on V_z will be limited.

Similarly, with this single-dose study, the mean AUC was approximately 20% lower for the morbidly obese subjects, but this difference was not statistically significant. These results are consistent with the demonstration that tedizolid clearance scales with weight^β such that the mean β value is close to 0.4. Body surface area is a function of height and weight, with a β exponent of 0.425 for the Du Bois-Du Bois equation and a value of 0.5 by Mosteller's adaptation (13, 14). An explanation of the mathematical similarity of ABW, LBW, and BSA as scalars has been detailed previously (15). Overall, body size explained less than 30% of the interindividual variability in tedizolid CL. A 3-fold increase in weight would be expected to result in a 55% (3^0.4) increase in tedizolid CL. Using a linear regression equation with a recognized poor coefficient of determination (Fig. 2), the estimated tedizolid CL rates are 5.13 liters/h and 7.01 liters/h in 65-kg and 130-kg individuals, respectively. Use of the power function (Fig. 2) yields a 31.6% difference in tedizolid CL estimates in comparing 130-kg and 65-kg individuals. The predicted 30% to 40% difference in CL between individuals with a 2-fold difference in body weight is small. Again, a lower tedizolid plasma AUC may be expected with fixed dosing in a 130-kg individual, but this difference is relatively small. Most importantly, the interindividual variability of tedizolid AUC was relatively small compared to that for linezolid (16). The mean (% coefficient of variation) AUC_0–∞ was 29.7 h · mg/liter (35.3%), with variability similar to those for the morbidly obese (29.3%) and nonobese (37.2%) groups.

The findings of this study are subject to the limitations of its design. The study evaluated healthy volunteers. The effects of acute illness caused by infection and the influence of comorbidities, such as diabetes, on the pharmacokinetics of tedizolid could not be determined. This study did not include elderly adults, and most of the subjects were female. Given that female subjects are shorter than males, on average, the weight of morbidly obese subjects was predominantly below 130 kg (to meet BMI criteria), which limits definitive extrapolations above this value. The pharmacokinetics of a single intravenous dose of tedizolid was studied, so the potential for pharmacokinetic differences in the oral bioavailability of tedizolid was not determined. Previous studies demonstrated limited systemic accumulation of tedizolid, suggesting similar concentration-time profiles for single and multiple doses. Given a convergence in the concentration-time profile at 24 h (Fig. 1) for both groups, the failure to study multiple doses is likely to be a minor limitation. Finally, additional samples should have been collected between the end of infusion and 2 h in order to characterize the pharmacokinetics of the parent compound (tedizolid phosphate).

Despite the aforementioned limitations, important conclusions can be drawn from this work. No significant differences in the plasma exposures (C_max, AUC_0–last, AUC_0–∞, and C_last) of tedizolid were observed among morbidly obese adults compared to age-, sex-, and ideal body weight-matched nonobese adults receiving a single 200-mg dose of intravenous tedizolid phosphate. Tedizolid CL and V_z correlated with body size, but the absolute differences in these parameters in comparing morbidly obese and nonobese subjects were small. Specifically, this analysis suggests that less than 30% of the interindividual variability in tedizolid CL is explained by body size. The results of previous analyses documenting a potentially lower response rate in obese diabetic patients are not likely a result of altered systemic exposure. This study confirms the findings of POP-PK analyses indicating that a tedizolid dose adjustment is not necessary for morbidly obese adults.