Tjokosela Tikiso 1, Helen Mcilleron 1, Nandita Sugandhi 2, Tim R. Cressey 3, Mark Mirochnick 4, Edmund V. Capparelli 5, Martina Penazzato 6, Paolo Denti 1
1. University of Cape Town, South Africa, 2. ICAP, Columbia University, New York, NY, USA., 3. Programme for HIV Prevention and Treatment, Chiang Mai, Thailand,4.Boston University School of Medicine MA, USA, 5.University of California San Diego, La Jolla, CA, USA.,6.WHO, Geneva, Switzerland.
Objectives: The widely accepted assumption underpinning the optimization of paediatric dosing regimens for anti-infectives is a similar PK/PD relationship to adults. The resulting strategy is therefore to target the same exposures proven to be effective in adults, while monitoring safety in children. The most common approach to devise a “best-guess” dosing approach before (or without) the availability of paediatric clinical data is to use the same mg/kg dose as in adults, despite strong evidence suggesting that this causes under-dosing, especially in smaller children. Allometric scaling and maturation have been shown to be much more reliable and accurate approaches to predict the effect of body size on PK parameters for most drugs, at least in children older >1 year. The possible reason for the persistent use of the constant mg/kg approach by clinicians is likely the fact that the nonlinear relationships underpinning allometry and maturation are more complex to understand and difficult to implement than the linear relationship implied by constant mg/kg dosing. To try and fill this gap between theory and implementation, the WHO Paediatric Anti-retroviral Working Group (PAWG) created a generic paediatric dosing tool [2] based on a spreadsheet specifically designed to aid the devising of paediatric dosing regimens from adult exposures in keeping with allometric scaling and maturation principles. While helpful, the current tool has limitations, in that it only compares overall AUC, it can still be confusing with its multiple sheets and complex formulae and has limited scope in creating detailed visualizations. The objective of this work is to improve the spreadsheet-based tool by re-implementing it with an open-source R shiny app and add additional functionalities.
Methods: The tool was redesigned with shiny dashboard package in R platform, it is a more interactive and intuitive data analysis tool than the MS Excel spreadsheet. It uses mlxR-Simulx [4] for simulating drug-specific population PK models and R Markdown for generating detailed reports. Similar to the spreadsheet implementation of the tool, the user can explore the expected overall exposure (AUC) for a generic drug by changing the strength of each tablet and number of tablets in each weight-band, either using the established WHO weight-bands or defining customised weight-bands. The R shiny implementation of the tool also includes the option of simulating entire PK profiles (concentration over time after dose) by using a drug-specific popPK model, which can include covariates and between-subject and –occasion variability. The popPK models can be loaded from the available library, or the user can provide custom popPK models implemented in mlxR-Simulx. The simulations can be performed using different in silico population data, such as the standard WHO/CDC growth charts, or a specific population (e.g. malnourished children, or custom data).
Results: The tool has been validated by assessing some of the current paediatric dosing regimens recommended by WHO for anti-retrovirals. The predicted exposures are closely in line with results from PK modelling and generally coherent with those obtained in confirmatory clinical trials. Most importantly, the doses predicted using the tool achieve exposures closer to the target and more even across weight-bands than those obtained with the constant mg/kg approach. The drug-specific models can simulate an entire PK profile, thus providing not only AUC, but also Cmax and Cmin. This allows the user to also consider potentially risky peak concentrations and explore alternative scenarios with different dosing frequency (e.g. twice daily).
Conclusions: We present a web-based implementation of the PAWG paediatric dosing tool, which is user-friendly and can also be accessed online with a web-browser. While the tool is only meant to be used as a first step in the design of clinical trials for paediatric dosing and not as a substitute to confirmatory studies, it represents a significant step forward with respect to using the constant mg/kg paradigm.
References:
[1] Dunne J, Rodriguez WJ, Murphy MD, et al. Extrapolation of Adult Data and Other Data in Pediatric Drug-Development Programs. Pediatrics. 2011;128:1242-1249. doi:10.1542/peds.2010-3487
[2] Denti P, Sugandhi N, Cressey TR, Mirochnick M, Capparelli E V., Penazzato M. An easy-to-use paediatric dosing tool: one mg/kg dose does not fit all. In: 24th Conference on Retroviruses and Opportunistic Infections. 2017. Abstract. 809
[3] R Core Team (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/
[4] Lixoft. mlxR – Simulx. Available at: http://simulx.webpopix.org/mlxr/. Accessed February 28, 2020.
Reference: PAGE () Abstr 9550 [www.page-meeting.org/?abstract=9550]
Poster: Methodology - Other topics