Evaluation of methods for estimating time to steady state with examples from phase 1 studies

An overview is provided of the methodologies used in determining the time to steady state for Phase 1 multiple dose studies. These methods include NOSTASOT (no-statistical-significance-of-trend), Helmert contrasts, spline (quadratic) regression, effective half life for accumulation, nonlinear mixed effects modeling, and Bayesian approach using Markov Chain Monte Carlo (MCMC) methods. For each methodology we describe its advantages and disadvantages. The first two methods do not require any distributional assumptions for the pharmacokinetic (PK) parameters and are limited to average assessment of steady state. Also spline regression which provides both average and individual assessment of time to steady state does not require any distributional assumptions for the PK parameters. On the other hand, nonlinear mixed effects modeling and Bayesian hierarchical modeling which allow for the estimation of both population and subject-specific estimates of time to steady state do require distributional assumptions on PK parameters. The current investigation presents eight case studies for which the time to steady state was assessed using the above mentioned methodologies. The time to steady state estimates obtained from nonlinear mixed effects modeling, Bayesian hierarchal approach, effective half life, and spline regression were generally similar.