Fixed dosing of intravenous tocilizumab in rheumatoid arthritis. Results from a population pharmacokinetic analysis

Abstract

Aims

Intravenous tocilizumab is currently dosed on body weight, although a weak correlation between body weight and clearance has been described. The aim of the study was to assess the current dosing strategy and provide a scientific rational for dosing using a modelling and simulation approach.

Methods

Serum concentrations and covariates were obtained from intravenous tocilizumab treated subjects at a dose of 4, 6 or 8 mg every 28 days. A population pharmacokinetic analysis was performed using nonlinear mixed effects modelling. The final model was used to simulate tocilizumab exposure to assess a dosing strategy based on body weight or fixed dosing, using as target a cumulative area under the curve at 24 weeks of treatment above 100 × 10³ μg h ml^–1.

Results

A one‐compartment disposition model with parallel linear and nonlinear elimination best described the concentration–time data. The typical population mean values for clearance, apparent volume of distribution, maximum elimination rate and Michaelis–Menten constant were 0.0104 l h^–1, 4.83 l, 0.239 mg h^–1 and 4.22 μg ml^–1, respectively. Interindividual variability was included for clearance (17.0%) and volume of distribution (30.8%). Significant covariates for clearance were patient body weight and C‐reactive protein serum levels. An estimated exponent for body weight of 0.360 confirms the weak relationship with tocilizumab clearance. Simulations demonstrate that patients with lower weights are at risk of underdosing if the weight‐based dosing approach is used. However, fixed‐dosing provides a more consistent drug exposure regardless of weight category.

Conclusions

Our study provides evidence to support fixed dosing of intravenous tocilizumab in rheumatoid arthritis patients since it reduces variability in tocilizumab exposure among weight categories compared to the current weight‐based dosing approach.

What is Already Known about this Subject

• Although intravenous tocilizumab effectiveness has been shown to be superior when administered at a dose of 8 mg kg^–1 compared to 4 mg kg^–1, recommended starting doses by agencies differ.

• Dosing is based on body weight even though body weight has shown a correlation of limited clinical relevance with drug clearance.

What this Study Adds

• This study identifies new sources of variability, such as inflammation patient status, and quantifies interpatient variability in intravenous tocilizumab pharmacokinetics.

• This study confirms that fixed dosing is appropriate in adult patients with rheumatoid arthritis reducing variability in tocilizumab exposure compared to current body size dosing strategy.

Introduction

Biological therapies are recommended for patients with rheumatoid arthritis (RA) with a nonsatisfactory response to synthetic conventional disease‐modifying antirheumatic drugs such as methotrexate, sulfasalazine and leflunomide.

Tocilizumab is a humanized monoclonal antibody that has demonstrated its efficacy in the treatment of RA by competitively inhibiting interleukin 6 binding to both its soluble and membrane receptors, thus preventing its proinflammatory activity 1, 2, 3, 4.

Currently, tocilizumab is available as intravenous (IV) infusion and subcutaneous injection. Different tocilizumab starting doses have been approved by the European and US regulatory agencies 5, 6. Furthermore, there is a trend in clinical practice to implement dose de‐escalation strategies in patients with RA who are in sustained remission, to improve the risk/benefit ratio and reduce costs 7, 8. Dose reductions are performed empirically mainly due to the lack of consistent pharmacokinetic (PK) studies that would help to assist in dosing recommendations. Moreover, recommended dosing regimens for IV administration are based on body weight whereas a fixed dosing is recommended for the subcutaneous administration. A fixed dosing approach is feasible for monoclonal antibodies since, in general, most of them are target specific, have a relatively large therapeutic window and body size is of minor influence on drug exposure.

To date, IV tocilizumab effectiveness has been shown to be superior when administered at a dose of 8 mg kg^–1 compared to 4 mg kg^–1 9, 10, 11, 12 and in those individuals who achieve a cumulative area under the curve (cAUC) at 24 weeks of treatment above 100 × 10³ μg h ml^–1 13.

Moreover, a population PK study 14 associated tocilizumab clearance (CL) with body surface area, showing that patients with a higher body surface area had a higher drug CL. In addition, body weight also affected CL. However, differences in body weight translated only into minor changes in CL, making it potentially clinically irrelevant. Therefore, a dosing approach based on a fixed (weight independent) tocilizumab dose is likely to be beneficial as it may show reduced PK variability and its preparation is much more convenient, cheaper and less‐error prone.

The aim of this study was to assess the current dosing strategy of IV tocilizumab in patients with RA and to provide a scientific rational for dosing using a modelling and simulation approach.

Methods

Patients

This study is a prospective, observational, single‐centre study conducted in the Arthritis Unit (Rheumatology Service) of the Hospital Clinic of Barcelona, a university tertiary hospital. The protocol (code HCB/2015/0533) was approved by the hospital ethics committee and was conducted in accordance with the Declaration of Helsinki and national and institutional standards. Before inclusion in the study, all patients provided written informed consent. Subjects eligible to participate in the study needed to be aged at least 18 years with diagnosis of RA, treated with IV tocilizumab and to have received a minimum of three doses of the study drug. Exclusion criteria were hypersensitivity to tocilizumab, pregnancy, active severe infections, hepatic enzyme alteration (>5 times the superior limit of normality), absolute neutrophil count <0.5 × 10⁹ l^–1 or a platelet count <50 × 10⁹ l^–1. Enrolled subjects received treatment with a 1–h infusion of IV tocilizumab at a dose range of 4, 6 or 8 mg kg^–1 every 28 days at the discretion of the treating physician.

Data

Demographic characteristics, such as age, sex, race, total body weight, height and smoking habit were recorded in the first evaluation. Clinical laboratory measurements including serum creatinine, creatinine clearance (estimated according to the Cockcroft–Gault formula 15), haematocrit, total serum proteins, albumin, high‐density lipoprotein–cholesterol, erythrocyte sedimentation rate (ESR) and C‐reactive protein (CRP) were documented at the moment of inclusion. Blood samples were collected to measure trough concentrations (just before the next drug dose) and, when possible, once a week until the next drug administration.

Serum was obtained and stored at –80°C until analysis. Tocilizumab concentrations were determined in the Immunology Department, CDB, Hospital Clinic, using the Lisa‐Tracker Tocilizumab Kit (Ref. LTT 002–96), ThearaDiag, France, following manufacturer's instructions. The limit of detection of the assay was 1 μg ml^–1 and the lower and upper limits of quantification were 1 and 50 μg ml^–1, respectively. Samples exceeding the upper limit of quantification were diluted 1:4 or 1:8.

Detection of anti‐tocilizumab antibodies (ADA) was performed using the Lisa‐Tracker Duo Tocilizumab Kit (Ref. LTT005), ThearaDiag, France. Owing to the interference of circulating tocilizumab, ADA could not be measured if tocilizumab concentrations exceeded 1 μg ml^–1. The assay detection limit was 5 ng ml^–1 and the lower and upper limits of quantification were 5 and 100 ng ml^–1, respectively. Calprotectin serum levels were also analysed in all samples using Calprotectin ELISA (ALP) Kit (Ref. CALP0170), CALPROLAB CALPRO AS, Norway) in accordance with the manufacturer's protocol.

Clinical efficacy was assessed at the visit of inclusion by a rheumatologist through the measurement of the different variables that compose the most used composite indexes: the Disease Activity Score in 28 joints (DAS28), the Simplified Disease Activity Index (SDAI) and the Clinical Disease Activity Index (CDAI). Therefore, the number of swollen and tender joint counts, patient and physician global assessment of disease activity and pain patient assessment were recorded.

PK analysis

Nonlinear mixed effects modeling was performed using NONMEM v.7.3 (Icon Development Solutions) 16, following a three‐step strategy: (i) basic population model selection; (ii) covariate selection; and (iii) model evaluation. The first‐order conditional estimation with interaction (FOCEi) method was used for parameter estimation. The program Pirana version 2.9.5 was used as an interface for NONMEM and for run deployment 17. Models of one and two compartments with first‐order elimination were explored during the structural model development. Interindividual variability 18 was evaluated for all PK parameters. Additive, proportional and combined error models were tested for residual variability on drug concentrations.

Goodness of fit for a given model was assessed by changes in the NONMEM minimum objective function value (equal to –2log likelihood) and plots of population and individual predicted concentrations vs. observed tocilizumab concentrations. The residual error components of models were assessed via scatter plots of conditional weighted residuals vs. predicted concentrations and time 19, 20. Changes in the standard error of parameter estimates (precision) was also assessed. The minimal value of the objective function provided by NONMEM was used to discriminate between hierarchical models. A P value of 0.05, corresponding with a change in objective function of 3.84 points with one degree of freedom was considered statistically significant. R version 3.2.2 and the package Xpose4 version 4.5.3 were used to guide further model building process 21, 22. Pearlspeaks‐NONMEM (PsN) version 5.24 was used for automation throughout the modelling process 23.

In the second step, baseline demographic covariates (age, sex, ethnicity, body weight, height and smoking), baseline laboratory covariates (creatinine clearance, haematocrit, total serum proteins, albumin and high‐density lipoprotein–cholesterol) and baseline biomarker covariates (rheumatoid factor, anti‐CCP, anti‐drug antibodies and serum calprotectin) were included in the covariate analysis. The influence of time‐varying covariates such as CRP and ESR were also assessed. Covariates were entered by the cumulative forward inclusion/backward elimination procedures in NONMEM. A covariate was retained if it led to a significantly improved fit (P < 0.05, likelihood ratio test). Biological plausibility of the covariate, graphical displays based on the agreement between the observed and predicted drug concentrations and the uniformity of the distribution of the residuals were also considered for covariate inclusion. The extent of Bayesian shrinkage in the PK parameters was evaluated for each parameter in the final model 24.

The internal evaluation of the PK model was assessed by graphical and statistical methods, including visual predictive checks 25. Bootstrap resampling technique (with replacement) was used to build confidence intervals (CIs) of PK parameter estimates 26. The final model was fitted to the replicate data sets (2000 data sets) and parameter estimates were obtained for each of them. Median and 95% CIs values of the bootstrap sample estimates were compared with those estimated from the original data. In case of wide CI values, a sensitivity analysis was conducted to assess the robustness of the estimation of the parameter.

Dosage regimen simulations

Monte Carlo simulations using the final population PK model of IV tocilizumab were performed to assess whether the body‐weight dosing or fixed dosing was more appropriate. Therefore, the cAUC at 24 weeks of treatment was calculated with the two approaches using a random uniform distribution with a body weight between 40 and 120 kg: (i) 1000 patients were simulated after IV tocilizumab was administered at the doses of 8, 6 and 4 mg kg^–1 every 28 days for 24 weeks; (ii) 1000 patients were simulated after IV tocilizumab was administered at the doses of 560, 420 and 280 mg every 28 days for 24 weeks. Fixed doses were calculated from the body‐weight dosing recommendations considering a standard subject of 70 kg. All simulations were performed without considering the influence of CRP on tocilizumab CL and in a situation of systemic inflammation (CRP value of 2.8 mg dl^–1, according to mean patients' CRP baseline values in tocilizumab clinical trials) 1, 2, 4.

Statistical analysis

Statistical analysis was performed using R version 3.2.2. 21 Means, medians, standard deviations (SDs), 95% confidence intervals (95% CIs), and quartiles were calculated for continuous variables. The Student t test was used for comparisons of normally distributed variables and the Mann–Whitney U test was used for comparisons of non‐normally distributed variables. Results are expressed as absolute and relative frequencies for categorical variables, and the chi‐square test was used to compare them. The significance level for all analyses was defined by a P value of <0.05.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY 27, and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 28.

Results

Patients and samples

Thirty‐five subjects were included in the study; 31 (88.6%) were women. Mean (± SD) age was 54.1 (± 12.3) years and mean (± SD) total body weight was 63 (± 13.8) kg. Twenty patients were receiving concomitant treatment with steroids and 24 with synthetic conventional disease‐modifying antirheumatic drugs (see Table 1 for further detail). Fifty‐four percent of the patients received a dosage of 8 mg kg^–1 of IV tocilizumab (n = 19). The remaining patients were on reduced‐dose regimens: 6 mg kg^–1 (n = 8) and 4 mg kg^–1 (n = 8). Detailed patient characteristics and laboratory parameters are shown in Table 1. According to DAS28 status, 74%, 11% and 14% of patients were in remission, low disease activity and moderate activity, respectively. The percentages according to SDAI and CDAI were 9%, 49%, 37% and 6%, and 11%, 46%, 26% and 17%, respectively, for remission, low, moderate and high disease activity.

Table 1.

Summary of continuous and categorical demographic data, laboratory and disease activity values at baseline

Covariate	Value
Number of patients (% women)	35 (88.6)
Continuous
Age, Mean ± SD	54.1 ± 12.3
Weight, Mean ± SD	63.5 ± 13.8
Height, Mean ± SD	161 ± 8.10
Disease duration in years, Median (range)	11.1 (2.90−48.5)
Days from last tocilizumab dose, Mean ± SD	30.5 ± 3.90
Categorical
Ethnicity, n (%)
White/Caucasian	28 (80.0)
Hispanic	6 (17.1)
Afro‐American	1 (2.90)
Smoking habit, n (%)
Active	4 (11.4)
Ex‐smoker	9 (25.7)
Non‐smoker	22 (62.9)
Erosive RA, n (%)	29 (82.9)
Positive autoantibody status, n (%)
RF	25 (71.4)
Anti‐CCP	27 (77.1)
Previous treatment with biologic agents, n (%)	23 (65.7)
Cotreatment with corticosteroids, n (%)	20 (57.1)
Cotreatment with DMARDs, n (%)	24 (68.6)
Methotrexate	18 (75.0)
Leflunomide	5 (20.8)
Sulfasalazine	1 (4.2)
Laboratory data, mean ± SD
Serum creatinine (mg ml –1 )	0.69 ± 0.29
Haemoglobin (g l –1 )	134 ± 14.0
Total protein (g l –1 )	68.1 ± 4.59
Albumin (g l –1 )	43.5 ± 1.98
HDL‐cholesterol (mg dl –1 )	61.8 ± 13.8
C‐reactive protein (mg dl –1 )	0.29 ± 0.76
Erythrocyte sedimentation rate (mm)	4.69 ± 6.95
Disease activity scores, mean ± SD
DAS28	2.20 ± 1.00
SDAI	12.2 ± 7.53
CDAI	11.9 ± 7.60

CCP, cyclic citrullinated peptide; CDAI, Clinical Disease Activity Index; DAS28, Disease Activity Score in 28 joints; DMARDs, disease‐modifying antirheumatic drugs; HDL, high‐density lipoprotein; RA, rheumatoid arthritis; RF, rheumatoid factor; SDAI, Simplified Disease Activity Index.

A total of 109 blood samples were obtained during the study period. Fifty‐four samples correspond to trough concentrations and the rest correspond to intermediate time points between two drug administrations (approximately at 7, 14 and 21 days after drug infusion). Sixteen patients participated with a single blood sample and the remaining 19 with an average of five samples per patient.

ADAs were tested in all samples with tocilizumab levels <1 μg ml^–1 (n = 17), but they were not detected in any of them.

Population PK model

Tocilizumab serum concentrations were best described by a one‐compartment disposition model with parallel first order (linear) and Michaelis–Menten (nonlinear) elimination kinetics. Interindividual variability was incorporated into the CL and volume of distribution (V) parameters. Residual variability was characterized by a combined error model with an additive part of 0.161 μg ml^–1 (expressed as the SD) and a proportional part of 25.5% (expressed as the coefficient of variation). The covariates selected to be included in the final model were the effect of total body weight and CRP levels on CL (P < 0.05). The PK model for CL was expressed as the following:

$$\mathit{CL}=0.0104\cdot {\left(\frac{\mathit{WT}}{62}\right)}^{0.360}\cdot \left(1+0.131\cdot \left(\mathit{PCR}-0.484\right)\right)$$

An increase in weight from 40 to 120 kg led to a 49% increase in CL and an increase of CRP levels from 0.1 to 20 mg dl^–1 led to a 275% increase in CL.

The diagnostic plots for the basic and final PK model indicated that the observed and predicted data were in good agreement with symmetrical distribution in the conditional weighted residual plots. The model parameters had reasonable levels of η‐shrinkage for CL (33.9%) and V (16.1%). An overview of the PK parameter estimates from the final population PK model is presented in Table 2.

Table 2.

Tocilizumab parameter estimates and bootstrap results for the final population pharmacokinetic model

Parameter	Final Model estimate	Standard error (% RSE), shrinkage [%]	Mean (95% CI) bootstrap result
Fixed effects
CL, l h –1	0.0104	24.0	0.0116 (0.00573–0.0226)
V, l	4.83	12.5	4.79 (4.00–5.63)
V M , mg h –1	0.239	46.4	0.337 (0.140–0.877)
K M , μg ml –1	4.22	78.7	9.72 (1.85–26.6)
Interindividual variability
VAR(η CL ), CV%	17.0	62.6 [33.9]	24.6 (0.17–42.1)
VAR(η V ), CV%	30.8	25.5 [16.1]	29.8 (12.9–41.3)
Covariate effects
CRP (mg dl –1 ) on CL	0.131	123	0.364 (0.00131–1.39)
WT (kg) on CL	0.360	111	0.599 (0.00360–1.85)
Residual error
Additive, μg ml –1	0.161	103.1	0.349 (0.0552–0.562)
Proportional, %	25.5	16.6	24.9 (15.1–33.3)

CL, clearance; CRP, C‐reactive protein; CV, coefficient of variation; K_M, Michaelis–Menten constant; V, central volume of distribution; V_M, maximum elimination rate; VAR, variance; WT, total body weight.

Model evaluation

The visual predictive checks indicated adequate goodness‐of‐fit (Figure 1). The statistical distributions of the parameter estimates obtained from the bootstrap analysis are shown in Table 2. The median values of the parameters estimated from the bootstrap analysis were in good agreement with the point estimates and the 95% CIs were reasonably narrow, demonstrating satisfactory precision, except for Michaelis–Menten constant (K_M) CI, which was expected to be wider since information on target binding was lacking. High standard error values for some of the parameter estimated can be explained by sparse data, a common situation in studies conducted in the clinical practice setting. Visual and numerical predictive checks demonstrated good predictive performance of the final PK model (Figure 1).

Figure 1.

Visual predictive check of the final pharmacokinetic model. Comparisons were performed between the 10th, 90th (dashed lines) and 50th (solid line) percentiles of the observed tocilizumab plasma concentrations (open circles) vs. time after dose (h) and the 80% confidence interval (shaded area) obtained from 1000 simulations

Dosage regimen simulation

Results from the performance of the two tested approaches (body weight and fixed dosing) are displayed in Table 3, with mean simulated values for cAUC, maximum concentration (C_max) and minimum concentration (C_min) for each of them. As expected, cAUC, C_max and C_min increased with increasing tocilizumab dose and the nonlinear CL of tocilizumab led to a greater than dose‐proportional increase in all values. Additionally, we observed a decrease in all PK measures of 19%, 4% and 54% for cAUC, C_max and C_min respectively, when simulations are performed with a CRP value of 2.8 mg dl^–1 (Table 4).

Table 3.

Intravenous tocilizumab steady state values for cumulative area under the curve, maximum and minimum concentration at 24 weeks of treatment for body‐weight dosing and fixed dosing (results from 1000 simulations) without considering the influence of C‐reactive protein

	Body‐weight dosing			Fixed dosing
	8 mg kg–1	6 mg kg–1	4 mg kg–1	560 mg	420 mg	280 mg
Steady state, mean (SD)
cAUC × 10 3 , μg h ml –1	253.8 (78.2)	176.9 (56.9)	103.9 (35.3)	220.2 (39.7)	151.9 (26.4)	87.8a (14.3)
Cmax, μg ml –1	155.6 (56.3)	113.6 (42.4)	72.9 (28.4)	135.5 (32.6)	98.6 (25.0)	63.2 (17.5)
Cmin, μg ml –1	16.6 (12.3)	9.3 (8.1)	3.4 (3.8)	13.9 (10.1)	7.3 (6.2)	2.4 (2.5)

cAUC, cumulative area under the curve; Cmax, maximum concentration; Cmin, minimum concentration; SD, standard deviation.

^aValue below the cAUC target, 100 × 10³ μg h ml^–1

Table 4.

Intravenous tocilizumab steady state values for cumulative area under the curve, maximum and minimum concentration at 24 weeks of treatment for body‐weight dosing and fixed dosing (results from 1000 simulations) considering the influence of C‐reactive protein with a value of 2.8 mg dl^–1

	Body‐weight dosing			Fixed dosing
	8 mg kg–1	6 mg kg–1	4 mg kg–1	560 mg	420 mg	280 mg
Steady state, mean (SD)
cAUC × 10 3 , μg h ml –1	201.8 (60.6)	142.1 (44.1)	85.4a (27.6)	175.8 (31.6)	122.9 (21.3)	73.0a (12.0)
Cmax, μg ml –1	146.9 (55.9)	108.5 (42.2)	71.0 (28.4)	128.2 (34.4)	94.6 (26.4)	61.9 (18.2)
Cmin, μg ml –1	7.9 (7.6)	4.2 (4.7)	1.5 (2.1)	6.6 (6.4)	3.3 (3.7)	1.1 (1.4)

cAUC, cumulative area under the curve; Cmax, maximum concentration; Cmin, minimum concentration; SD, standard deviation.

^aValue below the cAUC target: 100 × 10³ μg h ml^–1

Box plots showing the results from 1000 simulations from our final population PK model, without considering the influence of CRP, for body‐weight dosing and fixed‐dosing strategies are depicted in Figure 2. The percentage of patients attaining the target (defined as a cAUC at 24 weeks of treatment >100 × 10³ μg h ml^–1) 13 with each dosing strategy and considering the influence of CRP is provided in Figure 3.

Figure 2.

Box plot of cumulative area under the curve (cAUC) values at 24 weeks of treatment with intravenous tocilizumab at different doses along weight (WT) categories, from our population final PK model, without considering the effect of C‐reactive protein. Target is set in a cAUC values above 100 × 10³ μg h ml^–1. Median (percentile 50: q50) of the data set is represented with a solid line inside the box. Box edges are the third (q75) and first quartile (q25) for the superior and inferior, respectively. Superior and inferior limits (q90 and q10) are depicted with a thin horizontal line. Dots represent outliers. (a) 8 mg kg^–1; (b) 6 mg kg^–1; (c) 4 mg kg^–1; (d) 560 mg; (e) 420 mg; (f) 280 mg. (Results from 1000 simulations)

Figure 3.

Percentage of patients reaching the efficacy target (cumulative area under the curve at 24 weeks of treatment with intravenous tocilizumab above 100 × 10³ μg h ml^–1) with intravenous tocilizumab at different doses: high dose, 560 mg and 8 mg kg^–1; medium dose, 420 mg and 6 mg kg^–1; low dose, 280 mg and 4 mg kg^–1 along a weight (WT) range, using two different dosing approaches: fixed dosing (solid line) vs. body‐weight based dosing (dashed line). Results from 1000 simulations, without considering the effect of CRP (left column) and considering a CRP of 2.8 mg dl^–1 (right column)

The usual regimen of 8 mg kg^–1 every 28 days shows a target attainment in 99.8% of the subjects, being 1.9% of the subjects with extreme low weight (<50 kg) at risk of underdosing (mean cAUC ± SD: 142.6 ± 23.0). Patients from weights of 60–120 kg are far above the cAUC efficacy threshold. By contrast, using a fix dose of 560 mg of IV tocilizumab, the totality of individuals would be above the target regardless of their weight and showing a narrower range and more consistent drug exposure along weight groups.

With reduced doses of 6 mg kg^–1 of tocilizumab, a lower fraction (90.5%) of subjects would achieve the target compared with the full tocilizumab dose. Again, patients with lower weights are at risk of underdosing, with percentages of 32.6 and 20.6 if weight is between 40–49 kg and 50–59 kg, respectively. A comparable tocilizumab fix dose of 420 mg allows a more homogeneous distribution of patients reaching the therapeutic target (99.7%) with a maximum percentage of 1.5% of patients at risk of underdosing in extreme overweight patients (>110 kg).

Finally, with the reduced dose of tocilizumab of 4 mg kg^–1, 52.2% of the subjects would reach the target and >30% of individuals in the weight category between 40 and 89 kg would be underdosed. A fix dose of 280 mg provides a cAUC value above the target for 17.9% of the subjects, relatively independent of weight.

Simulating the same dosing approaches but considering a systemic inflammatory status with CRP levels of 2.8 mg dl^–1, showed a similar pattern across different body weight categories but with a lower percentage of patients reaching the PK target as expected (Figure 3, right column).

Discussion

Simulations confirm that the fixed‐dosing approach for IV tocilizumab reduces variability in tocilizumab exposure among weight categories compared to the current weight‐based dosing approach.

In this study, we used mixed effects modelling to characterize the PK of the drug and to investigate quantitative relationships between the PK parameters and physiological and/or demographic features of subjects with RA disease who were treated with IV tocilizumab. The serum concentration–time courses for the collected data were best described by an open one‐compartment PK model with parallel linear and nonlinear elimination. Although antibodies PK has usually been described using two‐compartment models, when the sampling strategy is insufficient, such as the lack of intense sampling just after the drug infusions as in our case, the peripheral compartment is not identifiable 29. However, the population estimates for CL and V were 0.0104 l h^–1 and 4.83 l, respectively, which are in close agreement with previously published values 14. Moreover, these data are in line with PK behaviour of other therapeutic monoclonal antibodies 30. Our work also showed considerable variability for CL and V as previously reported 14.

The development of ADA has been reported as a common cause for lower drug serum concentrations, and consequently, a loss of effectiveness after initial response with a monoclonal antibody 31, 32. However, response to tocilizumab has not been previously shown to be influenced by ADA 13. Our results support the low incidence of ADA formation in RA patients treated with IV tocilizumab. In fact, none of the serum samples analysed showed the presence of ADA in our population. For this reason, the influence of ADA on tocilizumab CL could not be tested.

Body weight has previously been identified as a covariate influencing the PK of some other monoclonal antibodies 1, 14. Approved tocilizumab dosing is based on body weight on the assumption that body‐weight dosing would provide a more consistent exposure across patients. This is under the assumption that exposure is linearly related to body weight 33. However, this relationship is far from linear for most monoclonal antibodies and as a consequence, this approach tends to overdose those patients with large body size while underdose patients with small body size 34, 35. This tendency is clearly seen in Figure 2a, 2b and 2c. On the contrary, fixed‐dosing may overdose patients with small body size and underdose patients with large body size.

Fixed dosing has already shown more consistent drug exposure than body‐weight dosing for some monoclonal antibodies. Wang et al. 34 claimed that the best dosing approach is basically dependent on the steepness of the relationship between CL and weight. Thus, taking into account that their relationship is generally expressed as a power function covariate model, for exponent values below 0.5, fixed‐dosing provides less variability in exposure. In our study, we estimated this exponent on 0.360, which is below this 0.5 threshold and, of similar magnitude as that of Frey et al. 14. Indeed, results from our simulations have shown that the distribution of cAUC values among weight groups after fixed dosing is much more uniform than after body‐weight dosing (Figure 2), thus supporting the use of fixed dosing. Moreover, a same trend in drug exposure was seen when simulations were conducted using the PK model of Frey et al., reconfirming that fixed dosing might be a better dosing strategy to achieve a more consistent drug exposure among all weight groups.

In addition, we found a significant effect of CRP on tocilizumab CL. To our knowledge, this is the first time that an inflammatory biomarker such as CRP is included in a PK model for a monoclonal antibody used in RA treatment 1. We believe it is a remarkable finding since higher values of CRP and ESR are common in clinical practice for those patients without a satisfactory response to treatment. As simulations results reflect, an increase in CRP value translates into a decrease in total drug exposure and thus, an increase in the percentage of patients at risk of being underdosed. Therefore, taking into account that high CRP values are usually seen at the beginning of the treatment, which would lead to a higher tocilizumab clearance, a starting dose of 8 mg kg^–1 (or the equivalent fixed dose of 560 mg) seems to be more reasonable than 4 mg kg^–1 (or 280 mg). Besides, it is probable that patients showing high values of CRP and not responding to treatment may be due to a higher drug CL, could benefit from a dose increment. However, according to Figure 3, a dose of 560 mg might be enough to reach target in the majority of the patients if CRP levels were 2.8 mg dl^–1. By contrast, drug CL would be lower in those patients showing low CRP values. Thus, a reduced tocilizumab dose might be enough for maintenance treatment.

In short, our study confirms the weak relationship between CL and body weight for tocilizumab and provides evidence to support fixed dosing of IV tocilizumab in RA patients. CRP levels have been identified as a clinically important covariate to affect tocilizumab CL. Routine measurement in the clinical setting may open up the possibility of its application in the future to predict treatment response and optimal dose.

Competing Interests

All authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no support from any organization for the submitted work, no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years and no other relationships or activities that could appear to have influenced the submitted work.

We would like to thank the nursing staff Fina Martin and Celia Saura (Rheumatology Department, Hospital Clinic of Barcelona) for their help in blood sampling, and Maria Torradeflot and Noemí de Moner (Immunology Department, CDB, Hospital Clinic of Barcelona) for sample processing.

This work was supported by a 1‐year grant of the Hospital Clinic of Barcelona (“Emili Letang Award – 2015”). C.B. holds an Obra Social “la Caixa” research grant – 2016.

Contributors

C.B., D.S and A.D.R.H. wrote the manuscript. D.S., R.S., V.R. and J.Y. designed the research project. C.B., V.R., R.S., M.P., J.Y. and D.S. performed the research. C.B., V.R., R.S., M.P., J.Y., A.H.M.V.S., A.D.R.H. and D.S. analysed the data. All authors revised the manuscript and accepted its form for submission.

Bastida, C. , Ruiz‐Esquide, V. , Pascal, M. , de Vries Schultink, A. H. M. , Yagüe, J. , Sanmartí, R. , Huitema, A. D. R. , and Soy, D. (2018) Fixed dosing of intravenous tocilizumab in rheumatoid arthritis. Results from a population pharmacokinetic analysis. Br J Clin Pharmacol, 84: 716–725. doi: 10.1111/bcp.13500.