Nicola Melillo (1), Silvia Grandoni (1), Nicola Cesari (2), Giandomenico Brogin (2), Paola Puccini (2), Paolo Magni (1)
(1) Università degli Studi di Pavia, Department of Electrical, Computer and Biomedical Engineering, Pavia, Italy; (2) Chiesi Farmaceutici S.p.A, Pharmacokinetic, Biochemistry and Metabolism Department, Parma, Italy.
Objectives: We performed a global sensitivity analysis (GSA) on an in-house physiologically based pulmonary absorption model for rats [1]. Physiological models are characterized by uncertainty and variability in the model parameters. It follows that the predicted metrics are uncertain (or variable) too [2]. As highlighted by the regulatory agencies, to improve the knowledge on the model, it is important to assess how much each parameter, with its variation, impacts on the model output variation. This can be achieved performing a GSA [3]. Here, we performed two types of GSA: inter-compounds and intra-compound. The aim of inter-compounds GSA is to understand what are the parameters that mostly influence the variability of the model predictions between different drugs. Instead, the aim of the intra-compound GSA is to understand how much the uncertainty associated with the parameters of a given drug impacts the model output uncertainty.
Methods: GSA methods consider each parameter Xi as a random variable, with associated a certain probability distribution [2]. Thus, the model output Y is a random variable too. For each Xi, variance based GSA derives two sensitivity indices from the decomposition of the variance (V) of Y: the main effect (Si) and total effect (ST,i). Both Si and ST,i are always included in [0,1]. The higher Si and ST,i are, the more Xi is important in explaining V(Y). Instead, ST,i=0 means that Xi is not influent on V(Y) [2].
In the inter-compounds GSA, drug related parameters are considered distributed into a range of values that includes all the compound of interest [4, 5]. All the parameters distributions were considered uniform within these ranges. The considered model outputs were the fraction absorbed (fa) and the lung-tissue AUC. We performed two GSA, one for highly soluble and the other one for poorly soluble compounds. The criterion used to divide compounds into these two classes resembles the one for orally administered compounds [6]. A dose number for inhaled compounds was defined (D0,inh) and it was used to distinguish highly and poorly soluble compounds.
We performed the intra-compound GSA on the absorption model coupled with a PBPK distribution model [7], for three internal compounds of interest. The parameter ranges of variation were defined equal to ±30% of the mean values, except for the active and passive permeabilities across lung tissues and for the blood to plasma ratio, that were considered equal to ±70% and ±10%, respectively. The considered outputs of interest were the lung tissue and plasma AUC.
Results: Concerning inter-compounds GSA, for highly soluble compounds, the parameter that mostly explain the fa variability is D0,inh (ST≈0.9) while for poorly soluble compounds are the mass median aerodynamic diameter (MMAD) and D0,inh (ST≈0.33 and 0.40, respectively). For lung AUC, the most important parameters for highly soluble compounds are the lung tissue binding (LTB) and both passive and active permeabilities (ST≈0.6, 0.6 and 0.2, respectively). For poorly soluble compounds, in addition to LTB and the permeabilities, D0,inh and MMAD (ST≈0.45 and 0.2, respectively) are also important. In the intra-compound GSA, for all the considered compounds, the lung AUC variance is mainly explained by the passive permeability variation (ST>0.5). The most important parameter for plasma AUC of the first internal compound is the extraction ratio (ST≈0.5), for the second one are the B:P, the dose and the rat weight (all ST≈0.25) and finally, for the third one are the dose, the rat weight and B:P (ST≈0.3, 0.2 and 0.25, respectively).
Conclusions: In the inter-compounds GSA, for highly soluble compounds, the most important parameters in explaining V(fa) are related with the dissolution process, while those for lung AUC variability, with drug retention in the lungs. With respect to highly soluble compounds, poorly soluble compounds have higher impact of parameters related with the dissolution process in explaining the lung AUC variability. In the intra-compound GSA was highlighted that the uncertainty related with the lung AUC is mainly explained by the passive permeability, while that of the plasma AUC by parameters related with drug distribution and metabolism. The inter-compounds GSA helps in understanding the model general behaviour in the whole parameters space, while the intra-compound GSA helps in identifying what parameters should be more precisely known to reduce the model output uncertainty.
References:
[1] Grandoni S, Cesari N, Melillo N, Brogin G, Puccini P, Magni P (2019) Development and evaluation of a PBPK model to study the pharmacokinetics of inhaled drugs in rats. PAGE 28, Submitted Abstract. Stockholm, Sweden.
[2] Saltelli A, Ratto M, Andres T, Campolongo F, Cariboni J, Gatelli D, Saisana M, Tarantola S (2008) Global Sensitivity Analysis. The Primer. John Wiley & Sons, Ltd
[3] CHMP (EMA) (2018) Guideline on the reporting of physiologically based pharmacokinetic (PBPK) modelling and simulation. Committee for Medicinal Products for Human Use (CHMP), European Medicines Agency (EMA), London, UK
[4] Melillo N, Aarons L, Magni P, Darwich AS (2019) Variance based global sensitivity analysis of physiologically based pharmacokinetic absorption models for BCS I–IV drugs. J Pharmacokinet Pharmacodyn 46:27–42 . doi: 10.1007/s10928-018-9615-8
[5] Melillo N, Aarons L, Magni P, Darwich AS (2018) Variance based Global Sensitivity Analysis of a Physiological Absorption model for compounds in different BCS classes. PAGE 27 (2018) Abstr 8559. Montreux, Switzerland
[6] Hastedt JE, Bäckman P, Clark AR, Doub W, Hickey A, Hochhaus G, Kuehl PJ, Lehr C-M, Mauser P, McConville J, Niven R, Sakagimi M, Weers JG (2016) Scope and relevance of a pulmonary biopharmaceutical classification system AAPS/FDA/USP Workshop March 16-17th, 2015 in Baltimore, MD. AAPS Open 2:1 . doi: 10.1186/s41120-015-0002-x
[7] Grandoni S, Cesari N, Brogin G, Puccini P, Magni P (2019) Building in-house PBPK modelling tools for oral drug administration from literature information. ADMET and DMPK 7:4–21 . doi: 10.5599/admet.638
Reference: PAGE 28 (2019) Abstr 9072 [www.page-meeting.org/?abstract=9072]
Poster: Drug/Disease Modelling - Absorption & PBPK