5613464 INTEGRATION OF PHARMACOKINETIC AND PHARMACODYNAMIC MODELS TO MULTI-CRITERIA DECISION ANALYSIS (MCDA) FOR PROSPECTIVE ASSESSMENT OF THE RISK-BENEFIT RATIO OF ORAL IRON CHELATORS IN PAEDIATRIC PATIENTS WITH TRANSFUSION-DEPENDENT HAEMOGLOBINOPATHIES

Abstract

Background: Current therapies for the treatment of thalassaemia and other hereditary rare haemoglobinopathies are symptomatic and aim at controlling or reducing iron overload by chelation of iron from a very early age. Iron chelation ultimately prevents complications or comorbidities that can significantly reduce patient survival. Given the processes underlying iron homeostasis and current understanding of the chelation mechanism of available chelating agents, it may be possible to implement alternative, more efficient approaches for evidence generation in support of the dose rationale and their use across different populations affected by transfusion-dependent haemoglobinopathies. Aim: This study aimed to characterise the benefit-risk balance (BRB) of deferiprone in different populations, including children < 2 years old and paediatric sickle cell patients, using a nonlinear mixed effects modelling approach in conjunction extrapolation concepts and multicriteria decision analysis (MCDA). Methods: Based on the underlying pharmacokinetic-pharmacodynamic relationships as well as on the incidence of adverse events, extrapolation concepts were applied to predict the BRB of treatment in young adults, paediatric patients younger than 2 years, and paediatric sickle cell disease patients receiving deferiprone. Using individual patient-level data from a randomised controlled clinical trial (DEEP-2, EudraCT 2012-000353-31), preference elicitation methods were applied to select and prioritise endpoints or outcomes to be used in the evaluation of the BRB across the populations of interest. Subsequently, model parameter estimates obtained from each model/endpoint were used in a set of clinical simulation scenarios in which the effect of chelation therapy was evaluated during 12 months in a virtual cohort of patients. Results: Five endpoints were selected and ranked by the experts involved in the preference elicitation process (Fig.1). ‘Absence or prevention of complications (diabetes/hypothiroidism)’ and ‘occurrence of agranulocytosis’ were the most important criteria, while ‘incidence of arthralgia’ and ‘occurrence of neutropenia’ were considered less important. Clinical trial simulations revealed that the overall benefit score of deferiprone was higher for the treatment of paediatric SCD patient compared with the reference population (i.e. DEEP-2 study). Small but statistically significant differences were observed for the use of DFP in children younger than 2 years old and adolescents or young adults. Table 1 summarises the clinical and demographic baseline characteristics of the DEEP-2 study and virtual cohort of patients included in the different simulation scenarios. Conclusion: This study has shown that MCDA can be utilised in combination with modelling and simulation principles to prospectively assess the effect of interventions in subgroups of patients for which clinical data generation is challenging or not feasible. It also offers an opportunity to evaluate the implications of dose adjustment as well as changes in the clinical management of these patients, providing a less empirical basis for dose selection and treatment choices in rare paediatric conditions such as transfusion-dependent haemoglobinopathies. Table 1 - Demographics of the populations in the MCDA analysis. Reference population reflects the actual patients in the DEEP-2 study, whilst scenarios reflect the baseline characteristics from the virtual patient population. Population Age (years) Weight (kg) Blood consumption (mL/kg/year) Baseline ferritin (ng/mL) Reference population (DEEP-2 study) 8.9 (2.7-17.2) 26.5 (12.1-53.5) 155.4 (17.6-404.3) 1991 (586.2-6488.7) Scenario 1: Adolescents and young adults 17.6 (11.3-33.8) 58 (29-83.6) 139.7 (28.5-347.6) 1933 (432-10106) Scenario 2: Children < 2 years old 1.3 (0.6-1.9) 6.3 (1.1-9.3) 162.2 (21-382.9) 726.3 (253.3-1757) Scenario 3: Paediatric sickle cell patients 14.5 (12.4-17.8) 39 (30-62.4) 139.5 (23.8-338.2) 1598.4 (476.3-8993.9)