Population pharmacokinetics and dos ing optimization...

1

Population pharmacokinetics and dosing optimization of vancomycin 1

in children with malignant hematological disease 2

3

Wei Zhao1,2 #, Daolun Zhang1, May Fakhoury1, Mony Fahd3, Frederique 4

Duquesne3, Thomas Storme4, André Baruchel3,5, Evelyne Jacqz-Aigrain1,2,5# 5

1. Department of Pediatric Pharmacology and Pharmacogenetics, Hôpital Robert Debré, 6

APHP, Paris, France 7

2. Clinical Investigation Center CIC9202, INSERM, Paris, France 8

3. AP-HP, Department of Pediatric Hemato-Oncology, Hôpital Robert Debré, Paris, 9

France 10

4. Department of Pharmacy, Hôpital Robert Debré, APHP, Paris, France 11

5. Université Paris Diderot, Sorbonne Paris Cité, Paris, France 12

13

Address for correspondence: 14

Doctor Wei Zhao 15

Professor Evelyne Jacqz-Aigrain, 16

Department of Pediatric Pharmacology and Pharmacogenetics, Hôpital Robert Debré 17

Clinical Investigation Center CIC9202, INSERM 18

48 Boulevard Sérurier, 75935 Paris Cedex 19, France 19

Phone: 0033 140033656 Fax: 0033 140035779 20

E-mail: 21

[email protected] 22

[email protected] 23

AAC Accepts, published online ahead of print on 24 March 2014Antimicrob. Agents Chemother. doi:10.1128/AAC.02564-13Copyright © 2014, American Society for Microbiology. All Rights Reserved.

on February 1, 2019 by guest

http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

2

Short title: Pharmacokinetics of vancomycin in children 24

Abstract: 249 words; Manuscript: 3442 words; References: 25; Figures: 4; Tables: 4. 25

Keywords: Vancomycin, Pharmacokinetics, Population pharmacokinetics, dosing 26

optimization, pediatrics, malignant hematological disease. 27

28


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

3

Background: Increased vancomycin dose has been proposed in adults with malignant 29

hematological disease. As pediatric data are limited, our aim was to evaluate the 30

population pharmacokinetics of vancomycin in order to define the appropriate dosing 31

regimen in children with malignant hematological disease. 32

Methods: Vancomycin concentrations were collected prospectively during therapeutic 33

monitoring. Population pharmacokinetic analysis was performed using NONMEM 34

software. 35

Results: Seventy children (age range: 0.3-17.7 years) were included. With current 36

recommended dosing regimen of 40-60 mg/kg/day, 53 children (76%) had sub-37

therapeutic steady-state trough concentrations (Css,min<10mg/L). A one-compartment 38

model with first-order elimination was developed. Systematic covariate analysis identified 39

that weight significantly influenced clearance (CL) and volume of distribution (V) with 40

power functions of 0.677 for CL and 0.838 for V. Vancomycin CL also significantly 41

increased with increasing of creatinine clearance and seems to be higher in children with 42

malignant hematological disease than in general pediatric population. The model was 43

validated internally. Its predictive performance was further confirmed in an external 44

validation by Bayesian estimation. The patient-tailored dosing regimen was developed 45

based on the final pharmacokinetic model, and showed that higher proportion of patients 46

reached the target Css,min compared to traditional mg/kg basis dose (60% versus 49%) 47

and reduced the risks associated with underdose or overdose. 48

Conclusion: This is the first population pharmacokinetic study of vancomycin in children 49

with malignant hematological disease. An optimized dosing regimen, taking into account 50


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

4

patient’s weight, creatinine clearance and susceptibility of the pathogens involved, could 51

routinely be used to individualize vancomycin therapy in this vulnerable population. 52

53


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

5

Introduction 54

Vancomycin is primarily effective against gram-positive cocci. Staphylococcus aureus 55

and Staphylococcus epidermidis, including both methicillin-susceptible (MSSA & MSSE) 56

or resistant-species (MRSA & MRSE), are usually sensitive to vancomycin. It is often 57

included in the initial empirical antibiotic therapy in patients with malignant 58

hematological disease (1). 59

Vancomycin is administered intravenously. It is 25-50% protein bound, mainly to 60

albumin and IgA, and is almost exclusively eliminated by the renal route. The 61

pharmacokinetic-pharmacodynamic breakpoint of vancomycin was defined as a ratio of 62

the area under the concentration-time curve in 24h (AUC0-24) over the minimum 63

inhibitory concentration (MIC) and is at least 400 h in adults with Staphylococcus aureus 64

pneumonia (2). 65

The pharmacokinetics of vancomycin has shown large inter-individual variability, 66

primarily linked to patients’ age, clinical condition and disease (3, 4). Patients in 67

oncology represent a critical population, in whom inadequate empirical antibacterial 68

therapy may result in infection-related morbidity and increased mortality. In addition, 69

pharmacokinetics often present different characteristics compared with non-oncology 70

patients (5-7), making optimization of dosing regimen essential. Vancomycin was 71

reported to have an increased clearance in adults with malignant hematological disease 72

compared with adults without cancer (8, 9). In children with cancer, pharmacokinetic 73

data are limited (10, 11) and the optimal dosing regimen remains undefined. Innovative 74

strategies, such as modeling and simulation approaches, were initiated to individualize 75

dose of vancomcyin in neonates and adults based on identified covariates (12). Thus, 76


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

6

the objectives of the present work were to evaluate the population pharmacokinetics of 77

vancomycin in children with malignant haematological disease and to optimize 78

vancomycin therapy based on defined pharmacokinetic-pharmacodynamic breakpoint. 79

80

Patients and Methods 81

Children with malignant hematological disease, receiving vancomycin as intravenous 82

infusion were included in the department of pediatric hematolo-oncology at Robert Debré 83

Hospital between 2010 and 2011 if at least one vancomycin serum concentration was 84

assayed for therapeutic drug monitoring (TDM). The following data, collected 85

prospectively by a trained research assistant, included: age, weight, serum creatinine 86

concentration, details of vancomycin administration (dose, infusion start and stop times), 87

type of hematological disease and bone marrow transplantation. Patients with incomplete 88

dosing information were excluded. This study was designed in accordance with legal 89

requirements and the Declaration of Helsinki, registered at the CNIL (Commission 90

Nationale Informatique et Liberté) and approved by the local research ethics committee 91

(CEERB Robert Debré Hospital, Paris France). 92

Dosing regimen and sampling 93

Vancomycin (Sandoz, Levallois-Perret, France) was administered as an intravenous 94

infusion over 60 min. The empirically initial dosing regimen is 40 to 60 mg/kg/day in 4 95

divided doses. Monitoring of vancomycin concentrations was performed in order to 96

maintain a steady-state trough concentration (Css,min) between 10 and 20 mg/L. 97


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

7

Assay of serum vancomycin and creatinine 98

The serum vancomycin concentrations were determined by a fluorescence polarization 99

immunoassay method (FPIA) using the Cobas Integra 400 plus system (Roche 100

Diagnostics, Meylan, France). The calibration curve ranged from 0.74 to 80 mg/L. The 101

accuracy and coefficients of variation (CVs) of controls (7.2, 21.4 and 35.6 mg/L) were 102

90%-110% and <3.9%, respectively. The lower limit of quantification (associated CV) 103

was 0.74 mg/L (CV<3.3%). Serum creatinine concentrations were measured by an 104

enzymatic method using the Advia 1800 chemistry system (Siemens Medical Solutions 105

Diagnostics, Puteaux, France). 106

Pharmacokinetic modeling 107

Pharmacokinetic analysis was carried out using the nonlinear mixed effects modelling 108

program NONMEM 7.2.0 (Icon Development Solutions, Ellicott City, MD, USA). First 109

order conditional estimation (FOCE) method with interaction option was used to 110

estimate the pharmacokinetic parameters and their variability. 111

Structure model 112

One or two-compartment open models with first order elimination were compared. The 113

basic model was evaluated through visual inspection of routine diagnostic plots. Inter-114

individual variability of the pharmacokinetic parameters was estimated using an 115

exponential model and was expressed as follows: 116

θ i= θ TV*eηi (equation 1) 117


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

8

where θ i represents the parameter value of the ith subject, θTV the typical value of the 118

parameter in the population and ηi the variability between subjects, which is assumed to 119

follow a normal distribution with a mean of zero and variance ω2. 120

Residual variability (additive, proportional, exponential or mixed) model was selected 121

according to improvement of the objective function value (OFV) and visual inspection of 122

routine diagnostic plots. 123

Covariate Analysis 124

The effects of age, weight, serum creatinine concentration, creatinine clearance and 125

type of hematological disease (leukemia or lymphoma) were investigated as potential 126

variables on pharmacokinetic parameters. The stepwise covariate modelling and 127

likelihood ratio test were used to test the effect of each variable. 128

The selection of covariates was based on a forward-backward process and biological 129

plausibility. During forward selection, a covariate was selected if a significant (p<0.05, χ2 130

distribution with one degree of freedom) decrease (reduction>3.84) in the OFV from the 131

basic model was obtained. At the end, all the significant covariates were added 132

simultaneously into a ‘full’ model. The importance of each covariate was re-assessed by 133

backward selection and independently removed from the full model if the increase in the 134

OFV was less than 7.88 (p<0.005, χ2 distribution). The resulting model was considered 135

the ‘final’ population pharmacokinetic model. 136

Model Validation 137


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

9

Model validation was based on graphical and statistical criteria. Goodness-of-fit plots, 138

including observed (DV) versus individual prediction (IPRED), DV versus population 139

prediction (PRED), conditional weighted residuals (CWRES) versus time and CWRES 140

versus PRED were used initially for diagnostic purposes (13). The stability and 141

performance of the final model were also assessed by means of a nonparametric 142

bootstrap with re-sampling and replacement. Re-sampling was repeated 500 times and 143

the values of estimated parameters from the bootstrap procedure were compared with 144

those estimated from the original data set. The entire procedure was performed in an 145

automated fashion, using PsN (14). The final model was also evaluated graphically and 146

statistically by visual predictive checks (VPC) and normalized prediction distribution 147

errors (NPDE) (15). One thousand datasets were simulated using the final population 148

model parameters. For the VPCs, the 5th, 50th and 95th percentile of the simulated 149

concentrations were processed by R, plotted against elapsed time and compared to 150

observed concentrations. For a model in which random effects are well estimated 151

approximately 90% of the observed data are expected to be within the 5th-95th prediction 152

interval. NPDE results were summarised graphically by default as provided by the 153

NPDE R package (v1.2) (16): (i) QQ-plot of the NPDE; (ii) histogram of the NPDE; 154

Given the objective of the analysis aiming to use of the final model for prediction 155

purposes, the predictive performance of developed model was further evaluated in an 156

independent group of children with malignant hematological disease. The individual 157

concentrations was predicted by Bayesian estimation (‘MAXEVAL = 0’ in the 158

ESTIMATION step) with NONMEM using the population pharmacokinetic parameters. 159


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

10

The predictive performance was evaluated by calculating the prediction error (PE) and 160

absolute prediction error (APE) using the following equations: 161

PE (%) = (Bayesian estimated concentration – measured concentration) / measured 162

concentration (equation 2) 163

APE (%) = ABS (Bayesian estimated concentration – measured concentration)/ 164

measured concentration (equation 3) 165

Dosing optimisation based on pharmacokinetic model 166

Monte Carlo simulations were performed using the parameter estimates obtained from 167

the final model in order to define optimal dosing regimen able to attain the target 168

AUC/MIC of 400 h in about 50% of patients, under the assumption of a comparable 169

safety profile. 170

Traditional pediatric dose (mg/kg basis) simulation approach 171

In this traditional approach, the paediatric dose of vancomycin was simulated on a 172

mg/kg basis according to different age group. Thus, various mg/kg dosing regimens (40, 173

50, 60, 70, 80, 90, 100 mg/kg) were simulated in each pediatric group: infants (28 days- 174

23 months), children (2-11 years) and adolescents (12-18 years). One thousand 175

simulations were performed using the original dataset, and AUC0-24 and Css,min at steady 176

state were calculated for each simulated patient. The target attainment rate was then 177

calculated for each dosing regimen to definite the optimal dose regimen in each 178

pediatric group. 179

Patient-tailored dose 180


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

11

Patient-tailored dose was assessed to evaluate the advantage of personalised therapy. 181

In this simulation scenario, the individual dose was calculated based on population 182

pharmacokinetic parameters and covariates in each patient as follow: 183

Optimized daily dosing i (mg/d) = target AUC0-24 × CLi (equation 4) 184

where CLi is calculated using the equation developed from the model; i stands for 185

individual. Target AUC0-24 was defined according to pharmacokinetic-pharmacodynamic 186

breakpoint of vancomycin: AUC0-24/MIC ≥400 h. For example, if the MIC was 1 mg/L, 187

the target AUC0-24 should be at least 400 mg•h/L. 188

The simulation process was similar to that described above for dose simulation on a 189

mg/kg basis. At the end, we compared the variability of AUC0-24 and Css,min between 190

mg/kg basis and patient-tailored dose. 191

192

Results 193

Serum concentrations of intravenous vancomycin infusion over 60 min were monitored 194

in 70 children (41 boys): mean (SD) age was 6.8 (4.8) (range 0.3 – 17.7) years and the 195

mean (SD) weight was 25.7 (15.5) (range 5.6 – 71.0) kg. A total of 98 vancomycin 196

concentrations were analyzed. A summary of patients’ demographic and clinical 197

characteristics is presented in table 1 198

Blood samples were drawn at the median of 54 hours after initiation of treatment. The 199

concentrations ranged from 1.8 to 27.3 mg/L. The number of patients in the 200


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

12

concentrations ranges (<5, 5-10, 10-15, >15 mg/L) were 12 (17%), 41 (59%), 11 (16%) 201

and 6 (9%), respectively. 202

Population pharmacokinetic Modelling 203

A total of 98 vancomycin concentrations were available for population modelling. Data 204

fitted a one-compartment model with first order elimination. Inter-individual variability 205

was best described by an exponential model and was then estimated for V and CL. 206

Residual variability was best described by a combined proportional and additive model. 207

The systematic covariate analysis identified bodyweight as the most important covariate 208

implemented on V, which caused a significant drop in the OFV of 35.5 points. For 209

clearance, bodyweight was also identified as significant covariate, causing a significant 210

drop in the OFV of 46.3 points. The model was further significantly improved by 211

introducing creatinine clearance (ΔOFV 26.3 points) on CL (Table 2). Therefore, the 212

influence of covariates on CL and V was retained in the model as follows: 213

CLi=CLref × (WTi / WTref) θ1× (CLCRi / CLCRref)

θ2 (equation 5) 214

Vi= Vref × (WTi / WTref) θ3 (equation 6) 215

where CLi and Vi are respectively the CL and V of the ith individual, WTi and CLCRi the 216

weight and creatinine clearance of the ith individual, WTref and CLCRref the reference 217

weight and creatinine clearance. The subscript “ref” indicates the individual with a 218

reference weight and creatinine clearance. In our study, the reference weight and 219

creatinine clearance were the median values of our population 20.2 kg and 191 mL/min. 220

The exponents were estimated by the model. 221


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

13

After incorporation of all significant covariates, inter-individual variability decreased from 222

61.6 to 34.8% for CL and from 117.9 to 77.0% for V. The shrinkage was 21% for CL and 223

23% for V. The final population pharmacokinetic parameters are given in table 3. 224

Model diagnostics showed acceptable goodness-of-fit criteria for the final model. As 225

shown in Figure 1A-D, population and individual predictions are acceptable. In addition, 226

the mean parameter estimates resulting from the bootstrap procedure very closely 227

agreed with the respective values from the final population model, indicating that the 228

final model is stable and can re-determine the estimates of population pharmacokinetic 229

parameters. The results of 500 bootstrap replicates are summarized in table 3. The 230

NPDE distribution and histogram indicates that the assumption of normal distribution of 231

the differences between individual predictions and observed data is acceptable (p=0.45 232

Shapiro-Wilk test of normality, Figure 1E-F). No trends were observed on the diagnostic 233

plots of NPDE versus time or predicted concentrations. The VPC (Figure 1G) shows 234

that observed concentrations were well predicted by the model (Exact Binomial Test, 235

11.2% out of limits observed, 95% confidence interval [5.7% – 19.2%]). Box plots of the 236

distributions of observed and predicted C0 from 1000 simulations derived from the final 237

model are presented are presented in Figure 1H. The median values of Css,min from the 238

observed data and the simulation as well as the interquartile range were similar, 239

indicating acceptable predictive capability of the final model. Figure 2 shows the 240

relationship between individual vancomycin clearance and covariates (body weight and 241

creatinine clearance). The typical CLs of patients weighting 20, 40 and 60 kg were 4.3, 242

6.9 and 9.1 L/h, respectively. 243


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

14

The performance of the developed model was further evaluated in an independent 244

group of 20 children with malignant hematological disease, with a mean (SD) age of 8.1 245

(4.1) (range 2.4 – 17.2) years, a mean (SD) weight of 28.8 (11.6) (range 13.9 – 62.0) kg 246

and a mean (SD) clearance creatinine of 191.3 (61.2) (range 84.0 – 350.7, Schwartz 247

formula) mL/min. 25 concentrations consisting of peak, trough and scavenged samples 248

were available and ranged from 4.1 to 67.2 mg/L. The Bayesian estimated 249

concentrations were highly correlated with measured concentrations (r2= 0.99). The 250

mean PE and APE were 1.0 % (5th-95th percentile: -4.7% – 8.7%) and 4.7%, 251

respectively, indicating a good predictive performance of developed model on new 252

patients. The estimates of population pharmacokinetic parameters did not change 253

obviously after inclusion of validation dataset into the model. 254

Dosing optimisation based on pharmacokinetic model 255

Traditional paediatric dose (mg/kg basis) 256

The target attainment rates as a function of dose and age groups for a standard MIC 257

susceptibility breakpoint of 1 mg/L is shown in Figure 3. The current recommended 258

dose of 60 mg/kg/day results in only 15% of infants and 24% of children and 259

adolescents achieving the target AUC/MIC. These simulated values were in agreement 260

with the observed values in the present study. In fact, 76% of our patients had risk of 261

underdose (Css,min <10 mg/L). To reach the target AUC/MIC of 400 h in about 50% of 262

patients, 90 mg/kg/day was required for infants and 80 mg/kg/day for children and 263

adolescents (Figure 3). The proportion of patients with risks of overdose (Css,min >20 264

mg/L) are 26% for the new proposed dosing regimen (Figure 3). 265


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

15

Patient-tailored dose 266

Patient-tailored dose was calculated for each patient based on equations 4 and 7: 267

CL= 4.37 × (Bodyweight / 20.2) 0.677 × (Creatinine clearance / 191) 1.03 (equation 7) 268

Where bodyweight is in kg and creatinine clearance is in mL/min (Schwartz formula). 269

The expected AUC0-24 and Css,min at steady state in simulated trials for infants receiving 270

90 mg/kg/day, and children and adolescents receiving 80 mg/kg/day in comparison to 271

patient-tailored dose are shown in Figure 4. The proportion of patients achieving target 272

Css,min (10-20 mg/L) is 60% using patient-tailored dose, which is higher than traditional 273

mg/kg basis dose (49%). The proportion of patients with risks of underdose (Css,min <10 274

mg/L) or overdose (Css,min >20 mg/L) are 20% and 20%, respectively, using patient-275

tailored dose, which is lower than traditional mg/kg basis dose (24% underdose and 276

26% overdose). 277

278

279

Discussion 280

In the present study, the population pharmacokinetics of vancomycin was evaluated for 281

the first time in children with malignant hematological disease. The high variability of 282

vancomycin concentrations emphasizes the need for dosing optimization and TDM-283

based dosage adaptation in this high-risk population. 284

The population pharmacokinetic model offered a practical tool to optimize vancomycin 285

dosing in children with malignant hematological disease. The rational dosing of 286


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

16

antimicrobials should take into account pharmacokinetics in the target population, the 287

relationship between drug exposure and outcome, and the susceptibility of the 288

pathogens involved. In children, the standard dose evaluation studies of antimicrobials 289

are usually based on a “non-selected” pediatric population and do not take into account 290

the potential impact of the disease and disease state which are the main factors that 291

ultimately influence drug exposure in the clinical setting (17). For the purposes of 292

comparison, the demographics and pharmacokinetic parameters of vancomycin 293

obtained from different pediatric studies (10, 18, 19) are summarized in table 4. The 294

magnitude of the differences provides strong argument for studying pharmacokinetics in 295

selected sub-groups of patients. Indeed, the estimated vancomycin CL in children with 296

malignant hematological disease was higher than that reported in general non-selected 297

pediatric populations. Such impact of hematological malignancies was previously 298

reported with vancomycin (8, 9) and additional antimicrobials in adult patients (5-7). 299

According to regulatory guidelines (20, 21), vancomycin is a good example of a drug for 300

which the modelling and simulation approach can be used to establish optimal dosage 301

recommendations in children. The pharmacokinetic-pharmacodynamic breakpoint of 302

vancomycin was defined as an AUC0-24/MIC ratio of at least 400 h in adults with 303

Staphylococcus aureus pneumonia (2), in which 76% (n=28) of methicillin-resistant 304

Staphylococcus aureus infected patients have a MIC of 0.5 mg/L, 24% (n=9) have a 305

MIC of 1mg/L. The pharmacokinetic-pharmacodynamic breakpoint is postulated to be 306

similar in children and is therefore used for dosing optimisation in children. 307

The current recommended pediatric dose of vancomycin 40-60 mg/kg/day (i.e. on a 308

classical mg/kg basis dose) and resulted in high risk of underdose in our target 309


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

17

population, as 76% patients had Css,min <10 mg/L. The dose simulation was in 310

agreement with this observation, and showed that the dose of 60 mg/kg/day resulted in 311

15% of infants and 24% of children and adolescents achieving the target AUC/MIC. 312

Therefore, in order to increase likelihood of treatment, vancomycin dose needs to be 313

increased in children with malignant haematological disease. In the present study, the 314

simulation approach demonstrated that the doses of 90 mg/kg/day for infants, and 80 315

mg/kg for children and adolescents, allow near to 50% patients to achieve the AUC/MIC 316

target, with the standard MIC susceptibility breakpoint of 1 mg/L, clinically validated 317

primarily in adult pneumonia studies in MRSA. However, according to recent data, there 318

is a trend to a decrease of vancomycin efficacy, linked to the increase of vancomycin 319

MIC during MRSA infection(22, 23). Indeed, to obtain similar efficacy, treatment of 320

MSRA with a higher MIC would require a higher AUC. Increase in vancomycin CL 321

together with creep in MIC values will require an increase of vancomycin daily dosage 322

to overcome the risk of therapeutic failure. However, up to now, the pharmacokinetic 323

parameters and safety of increased vancomycin dosing are not available. The dosing 324

simulation was based on the assumption of liner pharmacokinetics. Extrapolation of 325

dosing regimen outside of the observed range cannot preclude the possibility of non-326

liner pharmacokinetics and should be evaluated in further study. In addition, when 327

administering increased vancomycin dosing, the infusion-related adverse events (i.e. 328

red man syndrome) need to be evaluated and slowing the rate of infusion to at least 329

over two hours should be recommended. 330

One of the main advantages of modelling and simulation approaches is to optimize 331

personalized therapy. Therefore, in the second simulation scenario, we evaluated the 332


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

18

patient-tailored dose. Traditionally, pediatric dose is defined on a mg/kg basis according 333

to the different age groups, as we have done in the first simulation scenario. This 334

approach assumes an “average child” with an “average weight in each age group 335

(neonates, infants, children and adolescent) and a standard mg/kg dose is calculated 336

accordingly. However, both developmental factors and clinical conditions have a 337

profound impact on pharmacokinetics. Many demographic, biological and clinical 338

covariates are known to influence vancomycin pharmacokinetics. In our data, 339

vancomycin clearance increased with body weight and creatinine clearance, showing 340

that in children, the standard weight-based dosing (mg/kg) is not adapted to the whole 341

range of pediatric ages. This could be explained by the non-linear correlation between 342

bodyweight and developmental changes in vancomycin clearance (24). Furthermore, 343

the individual dose needs optimization according to abnormal renal function. Therefore, 344

weight and renal function have the most important impact in vancomycin 345

pharmacokinetics in children (3) providing the scientific basis for rational patient-tailored 346

dosing schemes. The simulation clearly supports the use of patient-tailored dose, which 347

showed a narrow spread AUC and Css,min compared with traditional mg/kg basis dose. 348

Patient-tailored dose resulted in higher proportion of patients within the target trough 349

concentrations, associate with lower risk of underdose or overdose. Individual dosage 350

prescription and adjustment based on these covariates will require a computer-based 351

tool. This practice has been set up in our NICU to individualize vancomycin dose in 352

neonates (25) and undoubtedly will be implemented for the treatment of pediatric 353

patients with “selected diseases”. 354


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

19

A limitation of our study was that the population pharmacokinetic model was developed 355

based on vancomycin TDM data, the prediction of maximum concentration at steady-356

state will likely shrink towards population mean values, reducing the ability to identify 357

“special” patient. Taking scavenged pharmacokinetic sample should improve individual 358

parameter estimation without increasing burden of pediatric clinical practice. 359

360

Conclusion 361

In the present study, we developed a population pharmacokinetic model of vancomycin 362

in children with malignant hematological disease. Vancomycin clearance was markedly 363

higher than that in children without cancer, with weight and creatinine clearance being 364

significant covariates. We have shown that the patient-tailored dose reduced variability 365

in vancomycin AUC and Css,min compared to the mg/kg basis dose, making the modeling 366

approach an important tool for dosing individualization. Explaining residual variability 367

requires the identification of additional covariates. A prospective study is warranted to 368

evaluate the potential clinical benefits and safety of this optimized dosing regimen. 369

370

371

Acknowledgments 372

We acknowledge the technicians (Christel Grondin, Michel Popon, Samira Benakouche 373

and Yves Médard) for technical support. This work was supported by Global Research 374

in Paediatrics – Network of Excellence (GRIP, EU-funded FP7 project, Grant 375


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

20

Agreement number 261060). The authors declare no conflict of interest related to this 376

work. 377

378


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

21

References: 379

1. Hughes WT, Armstrong D, Bodey GP, Bow EJ, Brown AE, Calandra T, Feld R, 380

Pizzo PA, Rolston KV, Shenep JL, Young LS. 2002. 2002 guidelines for the use of 381

antimicrobial agents in neutropenic patients with cancer. Clin. Infect. Dis. 34:730-751. 382

2. Moise-Broder PA, Forrest A, Birmingham MC, Schentag JJ. 2004. 383

Pharmacodynamics of vancomycin and other antimicrobials in patients with 384

staphylococcus aureus lower respiratory tract infections. Clin. Pharmacokinet. 43:925-385

942. 386

3. Marsot A, Boulamery A, Bruguerolle B, Simon N. 2012. Vancomycin: a review of 387

population pharmacokinetic analyses. Clin. Pharmacokinet. 51:1-13. 388

4. Zhao W, Kaguelidou F, Biran V, Zhang D, Allegaert K, Capparelli E, Holford N, 389

Kimura T, Lo YL, Peris JE, Thomson A, van den Anker JN, Fakhoury M, Jacqz-390

Aigrain E. 2013. External evaluation of population pharmacokinetic models of 391

vancomycin in neonates: The transferability of published models to different clinical 392

settings. Br. J. Clin. Pharmacol. 75:1068-1080. 393

5. Bubalo JS, Munar MY, Cherala G, Hayes-Lattin B, Maziarz R. 2009. Daptomycin 394

pharmacokinetics in adult oncology patients with neutropenic fever. Antimicrob. Agents. 395

Chemother. 53:428-434. 396

6. Romano S, Fdez de Gatta MM, Calvo MV, Caballero D, Dominguez-Gil A, Lanao 397

JM. 1999. Population pharmacokinetics of amikacin in patients with haematological 398

malignancies. J. Antimicrob. Chemother. 44:235-242. 399

7. Zeitany RG, El Saghir NS, Santhosh-Kumar CR, Sigmon MA. 1990. Increased 400

aminoglycoside dosage requirements in hematologic malignancy. Antimicrob. Agents. 401

Chemother. 34:702-708. 402


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

22

8. Buelga DS, del Mar Fernandez de Gatta M, Herrera EV, Dominguez-Gil A, García 403

MJ. 2005. Population pharmacokinetic analysis of vancomycin in patients with 404

hematological malignancies. Antimicrob. Agents. Chemother. 49:4934-4941. 405

9. Fernández de Gatta Mdel M, Santos Buelga D, Sánchez Navarro A, Dominguez-406

Gil A, García MJ. 2009. Vancomycin dosage optimization in patients with malignant 407

haematological disease by pharmacokinetic/pharmacodynamics analysis. Clin. 408

Pharmacokinet. 48:273-280. 409

10. Chang D. 1995. Influence of malignancy on the pharmacokinetics of vancomycin in 410

infants and children. Pediatr. Infect. Dis. J. 14:667–673. 411

11. Piro CC, Crossno CL, Collier A, Ho R, Koyama T, Frangoul H. 2009. Initial 412

vancomycin dosing in pediatric oncology and stem cell transplant patients. J. Pediatr. 413

Hematol. Oncol. 31:3-7. 414

12. Jacqz-Aigrain E, Zhao W, Sharland M, van den Anker JN. 2013. Use of 415

antibacterial agents in the neonate: 50 years of experience with vancomycin 416

administration. Semin. Fetal. Neonatal. Med. 18:28-34. 417

13. Hooker AC, Staatz CE, Karlsson MO. 2007. Conditional weighted residuals 418

(CWRES): a model diagnostic for the FOCE method. Pharm. Res. 24:2187-2197. 419

14. Lindbom L, Ribbing J, Jonsson EN. 2004. Perl-speaks-NONMEM (PsN)--a Perl 420

module for NONMEM related programming. Comput. Methods. Programs. Biomed. 421

75:85-94. 422

15. Brendel K, Comets E, Laffont C, Laveille C, Mentré F. 2006. Metrics for external 423

model evaluation with an application to the population pharmacokinetics of gliclazide. 424

Pharm. Res. 23:2036-2049. 425


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

23

16. Comets E, Brendel K, Mentré F. 2008. Computing normalised prediction 426

distribution errors to evaluate nonlinear mixed-effect models: the npde add-on package 427

for R. Comput. Methods. Programs. Biomed. 90:154-166. 428

17. Zhao W, Fakhoury M, Fila M, Baudouin V, Deschênes G, Jacqz-Aigrain E. 2012. 429

Individualization of valganciclovir prophylaxis for cytomegalovirus infection in pediatric 430

kidney transplant patients. Ther. Drug. Monit. 34:326-330. 431

18. Lamarre P, Lebel D, Ducharme MP. 2000. A population pharmacokinetic model for 432

vancomycin in pediatric patients and its predictive value in a naive population. 433

Antimicrob. Agents. Chemother. 44:278–282. 434

19. Wrishko RE, Levine M, Khoo D, Abbott P, Hamilton D. 2000. Vancomycin 435

pharmacokinetics and Bayesian estimation in pediatric patients. Ther. Drug. Monit. 436

22:522-531. 437

20. EMA. Points to consider on Pharmacokinetics and Pharmacodynamics in the 438

development of antibacterial medicinal products (Doc. Ref. CPMP/EWP/2655/99). 439

Available at http://www.ema.europa.eu/docs/en_GB 440

/document_library/Scientificguideline/2009 /09/WC500003420.pdf 441

21. FDA. Guidance for Industry: Exposure-Response Relationships — Study Design, 442

Data Analysis, and Regulatory Applications. Available at 443

http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guida444

nces/ucm072109.pdf 445

22. Ho PL, Lo PY, Chow KH, Lau EH, Lai EL, Cheng VC, Kao RY. 2010. Vancomycin 446

MIC creep in MRSA isolates from 1997 to 2008 in a healthcare region in Hong Kong. J. 447

Infect. 60:140-145. 448


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

24

23. Steinkraus G, White R, Fredrich L. 2007. Vancomycin MIC creep in non-449

vancomycin-intermediate Staphylococcus aureus (VISA), vancomycin-susceptible 450

clinical methicillin-resistant S. aureus (MRSA) blood isolates from 2001-2005. J. 451

Antimicrob. Chemo. 60:788-794. 452

24. Anderson BJ, Allegaert K, Van den Anker JN, Cossey V, Holford NH. 2007. 453

Vancomycin pharmacokinetics in preterm neonates and the prediction of adult 454

clearance. Br. J. Clin. Pharmacol. 63:75-84. 455

25. Zhao W, Lopez E, Biran V, Durrmeyer X, Fakhoury M, Jacqz-Aigrain E. 2013. 456

Vancomycin continuous infusion in neonates: therapeutic drug monitoring and dosing 457

optimization. Arch. Dis. Child. 98:449-53. 458


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

25

Figures legends 459

460

Figure 1: Model Evaluation 461

(A, B) Routine diagnostic goodness-of-fit plots: population predicted (PRED) versus 462

observed concentrations (DV). Individual predicted (IPRED) versus observed 463

concentrations (DV). 464

(C, D) Conditional weighted residuals (CWRES) versus time. Conditional weighted 465

residuals (CWRES) versus population predicted concentrations (PRED). 466

(E, F) Normalized Prediction Distribution Errors (NPDE): QQ-plot of the distribution of 467

the NPDE versus the theoretical N (0,1) distribution. Histogram of the distribution of the 468

NPDE, with the density of the standard Gaussian distribution overlaid 469

(G): Visual Predictive Check (VPC): Circles represent the observed, whilst dashed lines 470

depict the 5th and 95th percentiles of the simulated data. The solid lines indicate the 471

median obtained from the simulated data. 472

(H): Box plots of the distributions of observed and predicted Css,min from 1000 473

simulations. The bold horizontal bars in the middle show the median values of 474

vancomycin Css,min concentrations, while the outer boundaries of the boxes represent 475

the ranges of the 25th and the 75th percentiles (interquartile ranges). 476

477

Figure 2. The relationship between vancomycin clearance and covariates 478

A): Vancomycin clearance versus weight 479

B): Vancomycin clearance (normalized by size*) versus creatinine clearance 480

Dashed lines depict the dashed lines depict the typical covariate-CL relationship. 481

*Size=(Bodyweight / 20.2) 0.677 482


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

26

483

Figure 3. Target attainment rates and fraction of patients above maximum 484

recommended Css,min 485

486

The target attainment rates and fraction of patients above the maximum recommended 487

C0 for the 100 simulated trials is presented as a function of dose and age group. 488

AUC/MIC target is 400 h. MIC is 1 mg/L. The maximum recommended Css,min is 20 489

mg/L. 490

491

Figure 4. AUC and Css,min distribution. Simulated vancomycin AUC and Css,min 492

distribution in infants, and children and adolescents receiving mg/kg basis dose and 493

patient-tailored dose. 494


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

(A, B) Routine diagnostic goodness-of-fit plots: population predicted (PRED) versusobserved concentrations (DV). Individual predicted (IPRED) versus DV(C, D) Conditional weighted residuals (CWRES) versus time. CWRES versus PRED.(E F) Normalized Prediction Distribution Errors (NPDE): QQ-plot of the distribution of

Figure 1: Model Evaluation

(E, F) Normalized Prediction Distribution Errors (NPDE): QQ plot of the distribution ofthe NPDE versus the theoretical N (0,1) distribution. Histogram of the distribution ofthe NPDE, with the density of the standard Gaussian distribution overlaid(G): Visual Predictive Check (VPC): Circles represent the observed, whilst dashedlines depict the 5th and 95th percentiles of the simulated data. The solid lines indicatethe median obtained from the simulated data.(H): Box plots of the distributions of observed and predicted Css,min from 1000simulations. The bold horizontal bars in the middle show the median values ofvancomycin Css,min concentrations, while the outer boundaries of the boxes representthe ranges of the 25th and the 75th percentiles (interquartile ranges).


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

Figure 2. The relationship betweenvancomycin clearance and covariatesvancomycin clearance and covariatesA): Vancomycin clearance versus weightB): Vancomycin clearance (normalized by size*)versus creatinine clearance

Dashed lines depict the dashed lines depict thet i l i t CL l ti hitypical covariate-CL relationship.Size=(Bodyweight / 20.2) 0.677


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

Figure 3. Target attainment rates and fraction of patients above maximum recommended Css,min

The target attainment rates and fraction of patients above the maximum recommended Css,min for the 100 simulated trials is presented as a function of dose andsimulated trials is presented as a function of dose and age group. AUC/MIC target is 400 h. MIC is 1 mg/L. The maximum recommended Css,min is 20 mg/L.


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

Figure 4. AUC and Css,min distribution. Simulated vancomycin AUC and Css min distribution in infants, and children and adolescents ss,minreceiving mg/kg basis dose and patient-tailored dose.


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

1

Table 1: Baseline characteristics of 70 children 1

Number Mean SD Median Range

Patients 70

Gender 29 F/41 M

Bone Marrow Transplantation 25

Samples 98

Vancomycin dose (mg) 327 203 250 50 –950

Vancomycin dose (mg/kg) 13.0 3.4 13.6 7.0 – 31.5

Vancomycin concentration (mg/L) 9.4 5.3 8.1 1.8 – 27.3

Weight (kg) 25.7 15.5 20.2 5.6 – 71.0

Age (years) 6.8 4.8 5.6 0.3 – 17.7

Serum Creatinine (µmol/L) 32 17 30 10 – 141

Creatinine Clearance (mL/min)# 199.8 63.3 191.0 48.7- 457.0

Hematology disease

Acute Lymphoblastic Leukemia 40

Acute Myeloblastic Leukemia 17

Juvenile Myelomonocytic Leukemia 5

Lymphoma 5

Other 3

2

# Creatinine clearance was calculated by Schwartz formula. 3

4


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

2

Table 2: Covariate analysis 5

PK parameters

OFV ΔOFV

Structure model 398.6

Body weight CL 352.3 -46.3

Age CL 368.8 -29.8

Creatinine clearance CL 377.2 -21.4

Serum creatinine CL 397.0 -1.6

Type of hematological disease (leukemia or lymphoma)

CL 396.6 -2.0

Body weight, age and creatinine clearance

CL 323.6 -75.0

Age and creatinine clearance CL 351.3 -47.3

Body weight and creatinine clearance

CL 326.0 -72.6

Body weight V 363.2 -35.5

Age V 364.4 -34.2

Body weight and age V 357.9 -40.7

Allometric model (Body weight on CL and V with allometric scaling, and creatinine clearance on CL)

314.4 -84.2

Final model (Body weight on CL and V, and creatinine clearance on CL)

312.8 -85.8

6


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

3

Table 3: population pharmacokinetic parameters of vancomycin and bootstrap results 7

(n=500) 8

RF is the renal function; CLcr is creatinine clearance (mL/min), which was calculated by Schwartz 9

formula. 10

PK

parameters RSE(%) Bootstrap

Median 2.5th 97.5th

Volume of distribution (L) V

V=θ1×(Bodyweight / 20.2) θ2

θ1 119 13.4 116 81 157

θ2 0.838 25.1 0.803 0.401 1.32

Clearance (L/h) CL

CL=θ3×(Bodyweight / 20.2) θ4×RF

θ3 4.37 4.8 4.33 4.01 4.76

θ4 0.677 12.2 0.668 0.505 0.878

RF= (CLcr / 191) θ5

θ5 1.03 21.2 1.03 0.585 1.55

Inter-individual variability (%)

V 77.0 35.9 69.3 43.4 96.2

CL 34.8 22.9 32.7 22.2 39.8

Residual variability

Proportional (%) 5.3 79.2 6.0 2.0 16.0

Additive (mg/L) 1.17 26.6 1.15 0.68 1.54


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

4

Table 4. Pharmacokinetics of vancomycin in children 11

12

Pediatric patients

with vancomycin

TDM (ref 18)

Pediatric patients

with vancomycin

TDM (ref 19)

Children with

cancer (ref 10)

Pediatric patients

with vancomycin

TDM (ref 10)

Children with malignant

hematologic disease

(the present study)

Number of patients 78 6 33 31 70

Age (years) 7 # 6.9 6.8 (0.3-17.7)

Weight 25 # 21.2 25.7 (5.6-71.0)

CL (L/h/kg) 0.103 0.11 (0.08-0.14) 0.149 0.114 0.22 (0.04-0.73)

13

Values are expressed as mean (range) 14


http://aac.asm.org/

Dow

nloaded from

http://aac.asm.org/

Population pharmacokinetics and dos ing optimization...

Documents

Transcript of Population pharmacokinetics and dos ing optimization...