deciding on success criteria for predictability of...

DMD # 58099

1

Deciding on Success Criteria for Predictability of Pharmacokinetic

Parameters from In Vitro Studies: An Analysis Based on In Vivo

Observations

Khaled Abduljalil, Theresa Cain, Helen Humphries, Amin Rostami-Hodjegan

Simcyp Limited (a Certara company), Sheffield, S2 4SU, UK (K.A., T.C., H.H., A.R-H)

School of Pharmacy and Pharmaceutical Sciences, University of Manchester, Manchester,

UK (A.R-H)

DMD Fast Forward. Published on July 2, 2014 as doi:10.1124/dmd.114.058099

Copyright 2014 by the American Society for Pharmacology and Experimental Therapeutics.

This article has not been copyedited and formatted. The final version may differ from this version.DMD Fast Forward. Published on July 2, 2014 as DOI: 10.1124/dmd.114.058099

at ASPE

T Journals on July 12, 2020

dmd.aspetjournals.org

Dow

nloaded from

http://dmd.aspetjournals.org/

DMD # 58099

2

Running title: Prediction Success Criteria for Pharmacokinetic Parameters

Correspondence

Khaled Abduljalil

Simcyp Limited

Blades Enterprise Centre

John Street

Sheffield

S2 4SU

UK

Tel +44 (0) 114 292 2321

Fax +44 (0) 114 292 2333

Email: [email protected]

Text Pages: 26

Tables: 4

Figures: 5

References: 19

Abstract: 250

Introduction 616

Discussion 1061

Abbreviations: CL, systemic clearance (after intravenous administration); CYP,

Cytochrome-P450; IVIVE, in vitro to in vivo extrapolation; PD, pharmacodynamics; PK,

pharmacokinetics; Vss, volume of distribution at steady state (after intravenous

administration).


at ASPE



Dow

nloaded from


DMD # 58099

3

Abstract

Prediction accuracy of pharmacokinetic parameters is often assessed using prediction fold

error, i.e., being within 2, 3, or n-fold of observed values. However, published studies

disagree on which fold error represents an accurate prediction. In addition, ‘observed data’

from only one clinical study are often used as the gold standard for in-vitro to in-vivo

extrapolation (IVIVE) studies, despite data being subject to significant inter-study variability

and subjective selection from various available reports. The current study involved analysis

of published systemic clearance (CL) and volume of distribution (Vss) values taken from over

200 clinical studies. These parameters were obtained for 17 different drugs after intravenous

administration. Data were analysed with emphasis on the appropriateness to use a parameter

value from one particular clinical study to judge the performance of IVIVE., and the ability

of CL and Vss values obtained from one clinical study to ‘predict’ the same values obtained in

a different clinical study using the n-fold criteria for prediction accuracy. The 2-fold criteria

method was of interest as it is widely used in IVIVE predictions. The analysis shows that in

some cases the 2-fold criteria method is an unreasonable expectation when the observed data

are obtained from studies with small sample size. A more reasonable approach would allow

prediction criteria to include clinical study information such as sample size and the variance

of the parameter of interest. A method is proposed that allows the ‘success’ criteria to be

linked to the measure of variation in the observed value.


at ASPE



Dow

nloaded from


DMD # 58099

4

Introduction

In the top-down approach, where the model is derived from clinical data, pharmacometricians

use certain criteria for assessing pharmacokinetic predictions or model performance such as

goodness-of-fit plots. The prediction of interest is either concentration or response time

profiles. Such criteria are not used for bottom-up approach or IVIVE, where the predictions

of interest are is PK parameters, rather than the profiles. The justification of IVIVE prediction

is commonly performed by using the n-fold metric system and the success of IVIVE and

QSAR/QSPR prediction is usually assessed by determining the proportion of predictions

within a certain fold of a previously observed values. However, there is no specification of

which fold should be used and thus there is no consistency across different publications.

Published articles report their predictions within 1.5-fold (Han et al., 2013), 2-fold (Chen et

al., 2012), 3-fold (Gibson et al., 2009) and 5-fold (Gombar and Hall, 2013) of the observed

value. It is worth mentioning here that the term “observed values” refers to data collected for

use in the bottom-up approach to modelling and should be distinguished from that collected

for use in the top-down approach. In the bottom-up approach the term observed value refers

to the PK parameter values (eg., CL or Vss), while in the top-down approach the observed

values usually refers to the measured concentration in plasma. The aim of the top-down

approach is not to predict the PK/PD parameter itself, but to predict the impact of the PK

parameter on concentration, amount and or response in any biological matrix such as plasma,

urine, blood, or a tissue.

While the ‘observed data’ in the case of IVIVE predictions are the PK parameters, the values

considered to be gold standard data include uncertainties due to bias or imprecision in

calculating them from incomplete data but also because of the inherent variability. Reported

clinical PK parameters are subject to both inter- and intra-study variability, which stems from

different sources such as ethnicity, genetic variation in Cytochrome P-450 (CYP)


at ASPE



Dow

nloaded from


DMD # 58099

5

metabolizing enzymes (Tucker, 1994) such as CYP2D6 (Dorne et al., 2002; Abduljalil et al.,

2010; Thompson et al., 2011), CYP2C9 (Loebstein et al., 2001; Borgiani et al., 2007) and

CYP2C19 (Dorne et al., 2003; Crettol et al., 2005), co-medication, health status, different

assay and analysis, environmental factors and clinical settings of the study such as sample

size.

Inter-study variability in the clinical PK parameters may bias the assessment of IVIVE

prediction accuracy when predicted PK parameters are compared with data from only one

particular clinical study. It is also desirable if the predicted variability matches that of the

observed in the assessment of the predictability. These multiple sources of variability create a

challenge for comparing PK parameters between clinical studies themselves, and this

challenge increases significantly when one PK value is chosen to assess the performance of

IVIVE.

A significant amount of attention has been paid to prediction of PK parameters and their

dispersion from in vitro data (Howgate et al., 2006; Inoue et al., 2006; Jamei et al., 2009;

Cubitt et al., 2011), the accuracy of using PK parameters obtained from one clinical study to

‘predict’ the same parameter obtained in a different clinical study using the n-fold metric

system has not been investigated.

The aims of this paper are (i) to investigate whether parameters from one particular clinical

study can be used to judge the performance of IVIVE and if this is determined to be

appropriate , how to assess which clinical study the PK parameters should be taken from; (ii)

to investigate whether PK parameters obtained from one clinical study can ‘predict’ the same

values obtained in a different clinical study within a certain fold; (iii) to propose an improved

method for defining the prediction criteria that considers sample size and the variance around

the PK parameter of interest.


at ASPE



Dow

nloaded from


DMD # 58099

6

Materials and Methods

Clinical PK Parameter Data Collection

Data sources: Structured literature searches were carried out using Medline for the

parameters CL and Vss of 17 compounds (Table 1) , strictly after intravenous doses. The Vss

data after intravenous administration were collected for fewer drugs as the main focus was on

the ability of the n-fold metric system to predict parameters rather than the actual

predictability of the parameters. There were no criteria about compound selection other than

data availability. No language or date restriction was applied but article titles and abstracts

were screened to maintain the focus of the search upon these two parameters. Manual search

of reference lists from selected articles complemented the data collection process. Data were

extracted and entered into an Excel spreadsheet, which was subsequently checked prior to

analysis.

Data inclusion criteria were adult healthy Caucasian individuals. No restriction on gender or

maximum age was applied to the inclusion criteria as this is not applied during IVIVE. The

exclusion criteria were underlying health conditions known to affect the pharmacokinetic

parameters. For example, studies were excluded if they reported health conditions such as

renal insufficiency, smoking, cirrhosis, pregnancy and obesity. Studies that reported only

central tendency values without variability were also excluded from this study.

Clinical PK Parameter Data Analysis

Data analysis was performed using Microsoft Excel 2010. CL and Vss values were reported

with various units and obtained via compartmental and non-compartmental PK analysis. To

enable comparison between studies, all CL and Vss units were converted to L/h and L/kg,


at ASPE



Dow

nloaded from


DMD # 58099

7

respectively. A reference value of 70 kg body weight was assumed if the mean weight in the

original paper was not reported.

In this analysis, all reported CL values from different studies were put together for each

compound. A CL value among these values was randomly selected and assumed to be the

‘true’ value. This value was then compared with the remaining values that were considered to

be ‘predictions’ of the ‘true’ value. The procedure was repeated for each CL value within the

collected values for the drug X to allow each value equal chance to represent the true value,

and only once. This strategy of analysis was done for both CL and Vss separately and for each

of the 17 drugs in turn. ‘Predictions’ were plotted against the ‘true’ values for each parameter

for each compound, and the percentage of ‘predictions’ failing to be within 1.25, 1.5-, 2.0,

2.5- and 3-fold limits of the “true” value were calculated.

Hypothetical Data Simulation to Develop an Alternative Success Criteria

Trial mean Clearance values were simulated using the statistical software package R version

2.12 (www.r-project.org) to assess the suitability of the two-fold limits for varying sample

sizes and population CV% values. A population geometric mean CL value of 100 L/h was

assumed. This value was selected for convenience as comparing trials with the same actual

mean demonstrates the ability of the 2-fold prediction creiteria to accurately predict the

means. Trials were generated using this geometric mean, assuming sample sizes of between

5 and 20, and CVs between 10% and 150%. For each combination of sample size and CV%,

100,000 trials were simulated and the mean of each trial was then calculated.


at ASPE



Dow

nloaded from


DMD # 58099

8

The CL was assumed to have a lognormal distribution,CL~��log�100� , σ�, where σ is the

standard deviation of the data on the natural log scale (scale parameter of the lognormal

distribution) and is calculated from the CV% value using the equation

� � �� %��

�� 1� eq. 1

The trial mean Clearance values were simulated in R using the pre-defined function using for

generating values from a lognormal distribution. The percentage of trial means outside of the

two-fold limits of the population geometric mean were then calculated for each combination

of sample size and CV% value.

The trial means generated using a sample size of 10 were then investigated further by plotting

them on a graph and visually assessing their distribution relative to both the two-fold limits,

the 95% confidence intervals around the geometric mean CL and the new proposed metric

system which is the 99.998% confidence interval around the geometric mean CL. The 95%

geometric CIs are calculated using equation 2.

exp �ln�x�� 1.96 $ �

√% eq. 2

where ln�x�� is the natural logarithm of the mean value. The 99.998% geometric confidence

intervals were proposed as a new metric system in order to represent the majority of the

population distribution of the sample mean without having CIs between 0 and infinity. The

99.998% geometric CIs are calculated using equation 3.

exp �ln�x�� 4.26 $ �

√% eq. 3

The scale parameter σ value was calculated from the assumed CV% and sample size �.


at ASPE



Dow

nloaded from


DMD # 58099

9

These new metric using the 99.998% confidence intervals can be considered in terms of fold

limits instead of the standard two-fold limits. The two-fold limits can be written as 2��

0.5�� for geometric mean ��. In general the fold limits can be written as �� where the

predicted mean is accepted if it is within the upper ((��) and lower (��) fold limits of the

“true” geometric mean �� . These general upper and lower fold limits can be set equal to the

upper and lower 99.998% geometric mean CIs as shown in equations 4 and 5 respectively,

�� exp �ln�� 4.26 $ �

√�� eq. 4

and

�� exp �ln�� 4.26 $ �

√�� eq. 5

where σ is the sigma scale calculated (using equation 1) from the reported CV% and N is the

sample size of the reported mean in the clinical study. Equations 3 and 4 are considered as a

new fold metric approach to assess the ability of a parameter in the prediction of new

parameters.

Finally, this new proposed approach was used to evaluate the prediction of CL and Vss in

comparison to the 2-fold metric assessment as the cut-off. A successful and acceptable

prediction was considered if the predicted value was within the limit of the assessment

method of interest. The percentages of data within the 2-fold or 99.998% CIs were calculated

by counting how many inside this higher and lower boundaries as a percentage of the whole

data set for the parameter and compound of interest.


at ASPE



Dow

nloaded from


DMD # 58099

10

Results

Clinical PK Parameter Data Analysis

The data collected include wide spectral compounds that undergo extensive metabolism such

as midazolam or are eliminated mainly by renal excretion such as furosemide. The collected

data are presented in table 1 (see Supplemental Table A for the full list of the collected

references).

When comparing CL values from each study with the CL values in all other studies using the

2-fold metric for a given compound, it was observed that 13 of the 17 drugs had CL values

obtained from at least one clinical study that was outside the 2-fold limits of a CL value from

another study. . The percentage of mean CL values outside the 2-fold limits varied between

2% (lidocaine, theophylline and antipyrine) and 18% (digoxin). Only 4 of the 17 drugs had a

percentage error of ≥10% (propofol, midazolam, digoxin, alfentanil and lorazepam). This

percentage error was reduced to <5% for all compounds when the 3-fold accuracy criteria

was considered (Table 2).

Vss values after intravenous administration were available for 11 of the 17 drugs. Of the 11

compounds, 5 had mean Vss values that were outside the 2-fold limits of a value obtained

from another study. The range in percentage error for ‘predictions’ was between 2%

(lidocaine and diazepam) and 22% (midazolam). Only 2 of the 11 drugs had a percentage

error of ≥10% (midazolam and digoxin). This error was reduced to <5% for all compounds

when the 3-fold accuracy criteria was considered (Table 2)

Plots of each ‘true’ value for CL and Vss against all other ‘predicted’ values are presented in

Figures 1 and 2 respectively with the 2-fold system prediction limits. Several compounds

were randomly selected to represent- compounds of high, medium and low variability.


at ASPE



Dow

nloaded from


DMD # 58099

11

Hypothetical Data Simulation to Develop an Alternative Success Criteria

For each scale parameter value of the lognormal distribution and sample size, 100000 trials

were generated for an assumed CL population geometric mean of 100 L/h. The percentage of

trial means outside the 99.998% CIs of the population geometric mean are presented in

Supplemental Table B, by sample size and CV%. This is also presented graphically in Figure

3. The percentage of predicted means outside of the 99.998% CIs increases as the sample size

decreases and as the CV% of the PK parameter increases.

Figure 4 presents a plot of the simulated trial means using a sample size of 10 plotted against

the CV%. In this figure, three criteria; the two-fold limits and 95% and the new proposed

metric system using the 99.998% geometric CIs are plotted for comparison. As the scale

parameter (and therefore CV%) increases, the variability of the sample trial means increases,

as expected. For smaller scale parameter values (and therefore smaller CV%), the 2-fold

limits appear to be too wide to be used as prediction intervals and potentially allow values to

be accepted as good predictions which should not be, i.e. false positives. In contrast, for the

larger scale parameter values (and therefore larger CV%), not all trial means are within the 2-

fold limits and therefore values that are good predictions will not be accepted, i.e. false

negatives. The proposed method using the 99.998% CIs appear to include most sample trial

means and reduce the chance of either accepting clinically irrelevant values for the smaller

values of CV% or rejecting clinically relevant values for the larger values of CV%.

Comparison of 2-fold vs Alternative Success Criteria

A comparison between the 2-fold and alternative success criteria is presented in Table 3 and

Figure 5. The new proposed method using the 99.99% confidence intervals results in an


at ASPE



Dow

nloaded from


DMD # 58099

12

increased percentage of ‘predicted’ CL or Vss values outside of the proposed limits in

comparison to the 2-fold criteria as it is limited by the sample size and variability of the

compound. Using acetaminophen and propranolol as examples, it can be seen that the

boundaries of the new proposed method are contained within the 2-fold acceptance

boundaries (Figure 5).


at ASPE



Dow

nloaded from


DMD # 58099

13

Discussion

This study investigated the impact of in vivo variability of PK parameters and the metric

system commonly used to assess prediction accuracy for compounds with a wide range of

linear disposition properties.

There is evidence to suggest that in a clinical setting, the 2-fold prediction error in drug PK

parameters is acceptable for most drugs; however acceptance criteria will vary between

drugs. There is high variability in the PK parameters for some compounds, while a few have

relatively low variability (Table 1). For drugs with high variability, such as digoxin and

midazolam, the 2.5-fold metric system may be acceptable. For drugs with intermediate

variability like diazepam, the 2-fold system seems appropriate. For drugs with low variable

PK parameters, such as talinolol and acetaminophen, the 2-fold criteria boundaries are wide

and a tighter 1.5-fold could be appropriate.

In principle the prediction criteria described in the manuscript for clearance is also applicable

for other PK parameters such as volume of distribution. The predictability of volume of

distribution between studies in this analysis was better than that of clearance as there were

less variability in the case of distribution volume than in clearance. This is due to the fact that

clearance depends on area under the curve, while Vss calculation depends mainly on the first

few samples. A major limitation of the 2-fold criteria is that it handles data that comes from

different studies equally, irrespective of the sample size. For example it can accept a value

from studies with small sample sizes as good predictions if they are within 2-fold boundary

while reject the prediction of values from studies with high sample sizes if they are outside

the 2-fold boundary (see digoxin example in Figure 1). Previously, this system received

criticism in the field of drug interactions as it results in a potential bias toward successful

prediction at lower interaction levels and can bias any assessment of different DDI prediction


at ASPE



Dow

nloaded from


DMD # 58099

14

algorithms if databases contain a large proportion of interactions in the lower range of

interaction (Guest et al., 2011).

Simulation results presented in Figure 3 show that for a sample size of 10, the percentage of

means (of 100,000 trials) outside the 2-fold limits increases as the CV% value increases. This

is particularly the case as the sample size decreases as shown in Figure 4 which presents the

case for sample sizes between 5 and 20. The greatest percentage of trial means outside the

two-fold limits is 41%, for a sample size of 5 and a CV% of 150 (see Supplemental Table B).

However, for CV% of 10 and 20, all simulated trial means are within the two-fold limits.

This suggests that the percentage within the two-fold limits depends on both the sample size

and the CV% of the population. These findings show that the fold limits should be related to

the CV% of the population and the size of the sample used in the prediction.

Looking at the simulated dataset for a sample size of 10 (Figure 3), the 2-fold limits for

smaller values of CV% appear to be too wide to be used as prediction intervals, and could

allow the acceptance of a value that isn’t a true representation of that population, a high false

positive rate. Likewise, the number of simulated trial means outside the 2-fold limit for larger

CV% values suggests these limits are too small and could allow values not to be accepted as

representative of the population when they are true values, a high false negative rate. Both

sets of geometric CIs appear to be a similar shape to the distribution of simulated means,

suggesting the acceptance limits should be related to these CIs (Figure 3).

The new proposed method using the 99.998% CIs appear to include most of the simulated

trial means and if these limits were included instead of the two-fold limits it would reduce

both the false negative and false positive rates of prediction accuracy. If the 99.998% CIs


at ASPE



Dow

nloaded from


DMD # 58099

15

were used as the fold limits, it can be shown that the fold limits depend on both the sample

size and CV% as shown in Figure 4.

To apply the new method using 99.998% CIs, one need to calculate the sigma based on CV

and sample size from the reference in vivo study using equation 1 and substitute for sigma in

equations 4 and 5 which provide the lower and higher boundaries as mentioned in the

Methods section. If the predicted value from IVIVE was within this range, then the prediction

can be considered successful (i.e. not inconsistent with observed data). This helps to avoid

basing the decisions on goodness of prediction on the range obtained from small size studies.

The comparison between the 2-fold and new metric systems acceptance criteria was given in

Table 3 for all compounds used in the analysis and in Figure 4 for two selected compounds.

Taking the CL of propranolol as an example, the collected data gave an overall CL of

51.3±21.6 L/h (46%) mean±SD (CV%) from 12 clinical studies (see Supplemental Table A

for the list of these clinical studies) with a range of 30 - 76 L/h (Figure 5). The two lowest

and two highest CL values within this dataset were 30±5L/h (n=8), 35±4 (n=9), 71.4±8.4

(n=6) and 76±15 (n=12) L/h. It also shows that the CIs limits for accepting other reported

means as ‘predictions’ for these four mean CL values are 23.4 – 38.5, 29.8 – 41.1, 58.2 – 87.5

and 59.8 – 96.7 L/h, respectively.

The 2-fold limit for the smallest reported CL 30 L/h, is 15- 60L/h. This range will reject two

reported studies with mean values of 71.4 and 76 L/h and would accept the small unlikely

value of 15 L/h. However, if the highest reported CL value (76 L/h) was used as a reference

value, the 2-fold range of 38-152 rejects the two CL values of 30 and 35 L/h and accepts

values between 38 and 152 L/h. Neither a mean CL value of 152 L/h nor 15 L/h has been


at ASPE



Dow

nloaded from


DMD # 58099

16

reported. They are unlikely to be clinically relevant values for propranolol systemic CL in

healthy individuals either.

On the other hand, the new proposed method using 99.99% geometric confidence intervals

accepts a prediction range of 60.0 – 96.1 L/h for the observed CL value of 76 L/h with the SD

of 14.7 L/h that came from a study with a sample size of 12 individuals (Cheymol et al.,

1997). Likewise, it accepts a prediction range of 23.34 – 38.43L/h for the observed CL value

of 30L/h with the SD of 5L/h that came from a study with a sample size of 8 individuals

(Castleden and George, 1979). According to equations 3 and 4, the prediction of the new

proposed method will change if either the sample size or the dispersion parameters are

changed.

Limitations of this analysis are that, this analysis was carried out for selected studies based on

some selection criteria (see methods) and did not considered further subgrouping studies

according to the assay method or some demographics like gender, body weight, etc., as such

covariates are not available in all available publications. The collected data for those

compounds may not cover the actual variability of compounds under study. It should be

pointed out here that this work was done on the limited case of known linear drugs and the

situation might be in need of further assessment in the case of non-linear kinetics.

In conclusion, the arbitrary 2-fold system shows wide range of acceptance for low variable

drugs and vice versa. The discussed alternative prediction accuracy criteria will allow the

prediction to take into account clinical settings and parameter variability.


at ASPE



Dow

nloaded from


DMD # 58099

17

Acknowledgements

Assistance of Mr James Kay in gathering the data is greatly appreciated and Ms Eleanor

Savill for preparing the manuscript submission. We would also like to thank the MSc

Students in Modelling & Simulation course at the University of Sheffield and the University

of Manchester for their initial contributions to the analysis.

Authorship Contributions

Designed the study: Rostami, Abduljalil, Humphries

Data collection: Abduljalil, Humphries

Performed data analysis: Abduljalil, Cain

Wrote or contributed to the writing of the manuscript: Abduljalil, Cain, Humphries, Amin

Rostami-Hodjegan


at ASPE



Dow

nloaded from


DMD # 58099

18

References

Abduljalil K, Frank D, Gaedigk A, Klaassen T, Tomalik-Scharte D, Jetter A, Jaehde U,

Kirchheiner J and Fuhr U (2010) Assessment of activity levels for CYP2D6*1,

CYP2D6*2, and CYP2D6*41 genes by population pharmacokinetics of

dextromethorphan. Clin Pharmacol Ther 88:643-651.

Borgiani P, Ciccacci C, Forte V, Romano S, Federici G and Novelli G (2007) Allelic variants

in the CYP2C9 and VKORC1 loci and interindividual variability in the anticoagulant

dose effect of warfarin in Italians. Pharmacogenomics 8:1545-1550.

Castleden CM and George CF (1979) The effect of ageing on the hepatic clearance of

propranolol. Br J Clin Pharmacol 7:49-54.

Chen Y, Jin JY, Mukadam S, Malhi V and Kenny JR (2012) Application of IVIVE and

PBPK modeling in prospective prediction of clinical pharmacokinetics: strategy and

approach during the drug discovery phase with four case studies. Biopharm Drug

Dispos 33:85-98.

Cheymol G, Poirier JM, Carrupt PA, Testa B, Weissenburger J, Levron JC and Snoeck E

(1997) Pharmacokinetics of beta-adrenoceptor blockers in obese and normal

volunteers. Br J Clin Pharmacol 43:563-570.

Crettol S, Deglon JJ, Besson J, Croquette-Krokkar M, Gothuey I, Hammig R, Monnat M,

Huttemann H, Baumann P and Eap CB (2005) Methadone enantiomer plasma levels,

CYP2B6, CYP2C19, and CYP2C9 genotypes, and response to treatment. Clin

Pharmacol Ther 78:593-604.

Cubitt HE, Yeo KR, Howgate EM, Rostami-Hodjegan A and Barter ZE (2011) Sources of

interindividual variability in IVIVE of clearance: an investigation into the prediction

of benzodiazepine clearance using a mechanistic population-based pharmacokinetic

model. Xenobiotica 41:623-638.


at ASPE



Dow

nloaded from


DMD # 58099

19

Dorne JL, Walton K and Renwick AG (2003) Polymorphic CYP2C19 and N-acetylation:

human variability in kinetics and pathway-related uncertainty factors. Food Chem

Toxicol 41:225-245.

Dorne JL, Walton K, Slob W and Renwick AG (2002) Human variability in polymorphic

CYP2D6 metabolism: is the kinetic default uncertainty factor adequate? Food Chem

Toxicol 40:1633-1656.

Gibson CR, Bergman A, Lu P, Kesisoglou F, Denney WS and Mulrooney E (2009)

Prediction of Phase I single-dose pharmacokinetics using recombinant cytochromes

P450 and physiologically based modelling. Xenobiotica 39:637-648.

Gombar VK and Hall SD (2013) Quantitative structure-activity relationship models of

clinical pharmacokinetics: clearance and volume of distribution. J Chem Inf Model

53:948-957.

Guest EJ, Aarons L, Houston JB, Rostami-Hodjegan A and Galetin A (2011) Critique of the

two-fold measure of prediction success for ratios: application for the assessment of

drug-drug interactions. Drug Metab Dispos 39:170-173.

Han B, Mao J, Chien JY and Hall SD (2013) Optimization of drug-drug interaction study

design: comparison of minimal physiologically based pharmacokinetic models on

prediction of CYP3A inhibition by ketoconazole. Drug Metab Dispos 41:1329-1338.

Howgate EM, Rowland Yeo K, Proctor NJ, Tucker GT and Rostami-Hodjegan A (2006)

Prediction of in vivo drug clearance from in vitro data. I: impact of inter-individual

variability. Xenobiotica 36:473-497.

Inoue S, Howgate EM, Rowland-Yeo K, Shimada T, Yamazaki H, Tucker GT and Rostami-

Hodjegan A (2006) Prediction of in vivo drug clearance from in vitro data. II:

potential inter-ethnic differences. Xenobiotica 36:499-513.


at ASPE



Dow

nloaded from


DMD # 58099

20

Jamei M, Marciniak S, Feng K, Barnett A, Tucker G and Rostami-Hodjegan A (2009) The

Simcyp population-based ADME simulator. Expert Opin Drug Metab Toxicol 5:211-

223.

Loebstein R, Yonath H, Peleg D, Almog S, Rotenberg M, Lubetsky A, Roitelman J, Harats

D, Halkin H and Ezra D (2001) Interindividual variability in sensitivity to warfarin--

Nature or nurture? Clin Pharmacol Ther 70:159-164.

Thompson AM, Johnson A, Quinlan P, Hillman G, Fontecha M, Bray SE, Purdie CA, Jordan

LB, Ferraldeschi R, Latif A, Hadfield KD, Clarke RB, Ashcroft L, Evans DG, Howell

A, Nikoloff M, Lawrence J and Newman WG (2011) Comprehensive CYP2D6

genotype and adherence affect outcome in breast cancer patients treated with

tamoxifen monotherapy. Breast Cancer Res Treat 125:279-287.

Tucker GT (1994) Clinical implications of genetic polymorphism in drug metabolism. J

Pharm Pharmacol 46 Suppl 1:417-424.


at ASPE



Dow

nloaded from


DMD # 58099

21

Figure Legends

Figure 1: Analysis of whether a ‘true’ CL value from one clinical study can be used to

‘predict’ CL from other clinical studies within the 2-fold measure of prediction

accuracy given for Antipyrine, Digoxin and Talinolol as examples. Lines are the 2-fold

error around the identity line. The area of the bubble was set to the smallest size of the study

pairs. The smallest and largest bubble sizes are 6 and 16, respectively.

Figure 2: Analysis of whether a ‘true’ Vss value from one clinical study can be used to

‘predict’ Vss from other clinical studies within the 2-fold measure of prediction accuracy

given for Digoxin, Metronidazole and Midazolam as examples. Lines are the 2-fold error

around the identity line. The size of the bubble was set to the smallest size of the study pairs.

The smallest and largest bubble sizes are 7 and 96, respectively.

Figure 3: Simulated trial means for sample size 10 by lognormal distribution scale

parameter value, with two-fold limits and 95% and 99.99% geometric CIs.

Figure 4: Relationship between sample size, variance and percentage of predictions

outside the 99.998% CIs limit. As the sample size decreases and PK parameter variability

increases, the percentages of predicted means outside the two fold increases.

Figure 5: Comparison between the 2-fold vs the new prediction accuracy criteria for

propranolol (top plot) and acetaminophen (bottom plot) CL. Red dots are the range in

‘true’ CL values reported by the different clinical studies. The identity line represents the

same CL values plotted in numerical order (the ‘predicted’ CL values). Upper and lower 2-

fold acceptance boundaries are given as solid lines. The new method boundaries are given as

(-).


at ASPE



Dow

nloaded from


DMD # 58099

22

Table 1. Collected clinical data for CL (L/h) and Vss (L) after intravenous administration for 17 different compounds

Compound

CL Vss

No. of

Trials

Total no.

of Subjects

Mean CL (

range)a

Mean %CV

(range)

No. of

Trials

Total no. of

Subjects

Mean Vss

(range)a

Mean %CV

(range)

Talinolol 7 72 22.8 (19.1 -29.1) 16 (11 – 26) NA NA NA NA

Propranolol 12 102 52 (30 – 76) 17 (3 – 41) NA NA NA NA

Cisatracurium 12 210 20.5 (15.4 –

25.8)

18 (3 – 32) NA NA NA NA

Acetaminophen 11 146 19.4 (15.9 -

22.3)

19 (7 – 31) 12 146 0.99 (0.84 – 1.14) 19 (11 – 27)

Lidocaine 11 104 54 (31.5 – 77.7) 20 (6 – 32) 9 90 2.28 (1.47 – 3.34) 21 (9 – 28)

Propofol 15 305 127 (78 – 228) 21 (5 – 60) NA NA NA NA

Metronidazole 17 182 4.30 (2.10 –

6.43)

21 (6 – 43) 12 184 0.67 (0.51 – 0.80) 24 (7 – 50)

Theophylline 21 230 3.5 (2.1 – 4.7) 22 (6 – 83) NA NA NA NA

Ciprofloxacin 18 265 32.1 (18.3 –

52.2)

23 (10 – 60) 15 190 2.25 (1.82 – 2.91) 21 (13 – 39)

This article has not been copyedited and form

atted. The final version m

ay differ from this version.

DM

D Fast Forw

ard. Published on July 2, 2014 as DO

I: 10.1124/dmd.114.058099 at ASPET Journals on July 12, 2020 dmd.aspetjournals.org Downloaded from


DMD # 58099

23

Midazolam 33 570 28.6 (13.9 –

67.6)

23 (6 – 46) 21 251 1.30 (0.67 – 2.97) 17 (2 – 40)

Antipyrine 10 257 2.76 (1.85 –

3.91)

23 (6 – 48) 10 297 0.59 (0.48 – 0.72) 14 (4 – 33)

Digoxin 18 202 12.8 (4.5 – 24.5) 25 (13-47) 8 110 6.39 (3.45 – 8.41) 32 (11 – 64)

Furosemide 15 99 9.3 (6.2 – 16.1) 25 (7 – 50) 11 72 0.15 (0.11 – 0.21) 22 (6 – 58)

Thiopental 18 180 13.2 (9.2 -16.8) 29 (11 – 56) 12 117 2.05 (1.19 – 2.88) 37 (14 – 73)

Lorazepam 21 196 4.55 (2.52 –

8.32)

30 (10-49) 15 136 1.45 (1.25 – 1.85) 18 (8 – 34)

Diazepam 21 301 1.64 (1.01 –

2.69)

31 (11 – 51) 14 152 1.17 (0.82 – 1.83) 32 (11 – 57)

Alfentanil 14 193 18.5 (6.6 – 39.1) 40 (7 – 68) NA NA NA NA

a= arithmetic mean of clearance. NA= not applicable (data was not available)




DM

D Fast Forw




DMD # 58099

24

Table 2: Percentages of predictions outside the corresponding n-fold

Compound

CL Vss

1.25-fold 1.5-fold 2-fold 2.5-fold 3-fold 1.25-fold 1.5-fold 2-fold 2.5-fold 3-fold

Talinolol 24 4 0 0 0 NA NA NA NA NA

Propranolol 53 29 6 1 0 NA NA NA NA NA

Cisatracurium 36 8 0 0 0 NA NA NA NA NA

Acetaminophen 10 0 0 0 0 7 0 0 0 0

Lidocaine 48 20 2 0 0 54 22 2 0 0

Propofol 56 35 13 4 0 NA NA NA NA NA

Metronidazole 50 27 5 1 1 22 3 0 0 0

Theophylline 44 19 2 0 0 NA NA NA NA NA

Ciprofloxacin 43 17 3 1 0 21 3 0 0 0

Midazolam 56 32 11 4 2 73 48 22 11 5

Antipyrine 54 26 2 0 0 18 0 0 0 0

Digoxin 63 37 18 8 5 56 28 13 0 0

Furosemide 55 28 6 1 0 51 33 0 0 0




DM

D Fast Forw




DMD # 58099

25

Lorazepam 59 36 10 3 1 13 0 0 0 0

Diazepam 56 30 5 1 0 45 15 2 0 0

Thiopental 34 7 0 0 0 50 26 4 0 0

Alfentanil 68 45 16 8 3 NA NA NA NA NA

NA= not applicable (data was not available)




DM

D Fast Forw




DMD # 58099

26

Table 3: Percentages of CL and Vss predictions fall out of the corresponding limits 2-

fold vs the proposed method.

Compound

CL

2-fold Proposed system

Vss

2-fold Proposed system

Talinolol 0 27 NA NA

Propranolol 6 54 NA NA

Cisatracurium 0 44 NA NA

Acetaminophen 0 23 0 11

Lidocaine 2 39 2 42

Propofol 13 58 NA NA

Metronidazole 5 47 0 35

Theophylline 2 26 NA NA

Ciprofloxacin 3 38 0 21

Midazolam 11 54 22 59

Antipyrine 2 46 0 41

Digoxin 18 46 13 36

Furosemide 6 29 0 36

Thiopental 0 15 4 21

Lorazepam 10 37 0 16

Diazepam 5 32 2 19

Alfentanil 16 38 NA NA

NA= not applicable (data was not available).


at ASPE



Dow

nloaded from



at ASPE



Dow

nloaded from


deciding on success criteria for predictability of...

Documents