Felix Stader, Marco Siccardi, Melissa Penny, Hannah Kinvig, Manuel Battegay & Catia Marzolini
University Hospital Basel & University of Basel, Switzerland
Objectives: The aim of this work was to develop and verify a population database for people above the age of 65 years with necessary system parameters to inform physiologically-based pharmacokinetic (PBPK) models.
Methods: A structured literature search was performed to screen for age-dependency of anatomical, physiological and biological parameters required to inform a PBPK model. Included parameters were body height, body weight, organ weights, cardiac output, regional blood flows, plasma binding protein concentration and glomerular filtration rate. Abstracts were screened, and studies included if the study population was Caucasian (being defined broadly and included Europeans, North Americans and Australians), at least age was reported in addition to the parameter of interest, and individuals were healthy, or the disease was deemed unlikely to affect parameters of interest. Data were separated into a development and verification dataset. Studies in the development dataset had to report sex, body height, body weight, ethnicity and location of the study in addition to age as necessary covariates to be able to describe correlations between system parameters. Otherwise, studies with less reported covariates were used in the verification dataset. Age-dependency could be estimated if at least three different studies reported the parameter of interest with one value in each age decade. Otherwise, age-dependent effects could not be investigated, and similar values were assumed for elderly as defined previously in young subjects.
Linear regression was performed to derive descriptive, continuous equations from 20 to 99 years for the parameter of interest considering the independent covariates age, sex, anthropometric parameters, study location and publication year. Several data transformations, such as log-transformation, were investigated during regression analysis. Covariates with a p-value below 0.01 were considered as significant. The regression with the smallest sum of residuals was used. The performance of each derived equation was checked with an independent verification dataset. Variability for each parameter was calculated as the weighted coefficient of variance of the development dataset.
Results:
A total of 362 studies were found and finally 318 studies were included in the analysis adequately describing 44 out of 60 model parameters. Rich data were found for anthropometric parameters and for some organs, e.g. kidney and liver. Data for some regional blood flows, such as to the bone, and in general composition of tissues, were difficult to obtain from the literature and values for young subjects were required to be used. For tissue composition, the predicted sum of water and cell mass for all tissues assuming age-independent fractions coupled with age-dynamic tissue volumes were in accordance to reported total body water and total body cell mass showing the plausibility of used assumptions.
The developed population has been implemented in Matlab® and 1000 virtual men and women have been created. The estimated body height, body weight, organ weight, blood flows, plasma-binding protein concentration and glomerular filtration rate were in accordance to the independent verification dataset, demonstrating robustness of the developed population.
Conclusions: The developed population database for aging subjects can be implemented into existing PBPK frameworks and could subsequently allow the prediction of drug kinetics and drug-drug interactions in elderly.
Although including data for centenarians, most of the data were found from studies up to the mid-eighties identifying a knowledge gap for older individuals. It is worthwhile mentioning that for some parameters like blood weight, only data from the United States and studies published 50 years ago were found. The impact of location and publication year has been checked on the well described anthropometric parameters. Publication year was found to be a significant covariate for body weight, explained by a significant increase in body weight of adults over the last decade. This argues for the necessity to constantly update PBPK models and include body weight as a necessary covariate in the development dataset.
The developed repository provides continuous functions to describe anatomical and physiological system parameters from the age of 20 to 99 years, considers population variability and includes studies on the age-dependency of metabolising enzymes and drug transporters.
Reference: PAGE 27 (2018) Abstr 8461 [www.page-meeting.org/?abstract=8461]
Poster: Drug/Disease Modelling - Other Topics