Sample size calculations in multiple sclerosis using pharmacometrics methodology: comparison of a composite score continuous modeling and Item Response Theory approach
Ana Kalezic (1), Radojka Savic (2), Alain Munafo (3), Elodie Plan(1), Mats O Karlsson (1)
(1) Dept of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden (2) Dept of Bioengineering and Therapeutic Sciences, University of California San Francisco, San Francisco, USA (3) Merck Institute for Pharmacometrics, Merck Serono, Lausanne, Switzerland* *Part of Merck Serono S.A. Coinsins, Switzerland, an affiliate of Merck KGaA, Darmstadt, Germany.
Objectives: Clinical trials in multiple sclerosis (MS) therapeutic area are particularly long due to the variable and slowly progressive nature of the disease. Phase III trials are often conducted for over two years and frequently include more than a thousand patients. Therefore, increased efficiency would be valuable. This analysis aims to demonstrate the application of pharmacometric methods for power/sample size calculation based on Expanded Disability Status Score (EDSS) [1], a widely used measure of disease disability in MS, as efficacy endpoint.
Methods: Clinical trial simulations were used to compare the power to detect the drug effect for two Non-Linear Mixed Effect (NLME) models previously developed for EDSS. One is treating EDSS as a continuous, composite variable [2] and the second one is using Item Response Theory (IRT) methodology [3]. The IRT model was used for simulations according to predefined study design. Study design was a 96-week Phase III clinical study with relapsing-remitting MS, where patients received active treatment or placebo, and EDSS assessment was conducted every 12 weeks during the study. These scenarios were explored: that the drug has both offset and disease modifying effects or one of these alone.
The simulated data were subsequently analyzed using a continuous composite and an IRT model. The power calculations were performed using Monte Carlo Mapped Power method [4], implemented in PsN software. All simulations were performed and models fitted using NONMEM 7.2.
Results: In the current example, we show that using the IRT model requires 81 patients, compared to 131 with a continuous composite model to achieve 80% power to detect a dual drug effect. It represents 40% reduction in sample size. In case of offset and disease modifying drug effect alone, 20% and 40% fewer patients were necessary using IRT model, respectively.
Conclusions: Taking a model-based approach offers an opportunity to gain efficiency in clinical trials. In the current example, the IRT model indicated overall a need for lower sample size to detect the drug effect compared to continuous composite model regardless whether the drug effect was an offset, disease modifying, or both.
This is in line with previous findings that IRT increases precision in predictions and power to detect drug effects and linkage to biomarkers [5, 6]. Therefore, these NLME models could be used as a support for MS clinical trials design and analysis.
Acknowledgement: This work was supported by the DDMoRe project.
References:
[1] Kurtzke JF, Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS), Neurology 33 (11):1444-52, 1983
[2] Savic R, Munafo A, Karlsson MO. Population pharmacodynamics of cladribine tablets therapy in patients with multiple sclerosis: relationship between magnetic resonance imaging and clinical outcomes, Oral presentation at PAGE 2011, Abstract 2282
[3] Kalezic A, Savic R, Munafo A, Ueckert S, Karlsson MO. Application of item response theory to modelling of EDSS in multiple sclerosis, Poster at PAGE 2013, Abstract 2903
[4] Vong C, Bergstrand M, Nyberg J, Karlsson MO. Rapid sample size calculations for a defined likelihood ratio test-based power in mixed effects models, AAPS J 14(2):176-86, 2012
[5] Ueckert S, Plan EL, Ito K, Karlsson MO, Corrigan B, Hooker AC. Improved Utilization of ADAS-Cog Assessment Data Through Item Response Theory Based Pharmacometric Modeling; the Alzheimer’s Disease Neuroimaging Initiative, Pharm Res. 2014 Mar 5.
[6] Balsis S, Unger A, Benge J, Geraci L, Doody R. Gaining precision on the Alzheimer’s disease assessment scale-cognitive: A comparison of item response theory-based scores and total scores. Alzheimers Dement 2012; 8:288-294
