Glutathione peroxidase (GPx) is a key enzyme in the removal of hydrogen peroxide in cells and tissues.  It is widely believed that the rate of reduction of H₂O₂ to water by GPx is independent of the glutathione (GSH) concentration in typical in vivo settings, where [H₂O₂] < [GPx] < [GSH].  Therefore, the ability of GPx to remove H₂O₂ is indirectly estimated, commonly, by measuring a parameter related to the total activity of the enzyme.  However, a recent observation in an animal model indicated that the rate of tumor growth was a direct linear function of the effective GPx activity, i.e. [GPx]×[GSH], suggesting that the in vivo removal of H₂O₂ by GPx is governed by kinetics different than those observed in dilute, homogenous buffer solutions.  To understand this observation, a mathematical analysis of a kinetic model for the removal of H₂O₂ by GPx was undertaken to determine how the reaction species (H₂O₂, GSH, and GPx) influence the rate of removal of H₂O₂.  The mathematical kinetic model showed that when initial [GPx] is higher than [H₂O₂], the rate of removal of H₂O₂ is a function of neither [GPx] nor [GSH].  When [GPx] and [GSH] are relatively high, the overall time constant for removal of H₂O₂ shows some dependencies on either [GPx] or [GSH].  However, at lower [GPx] and [GSH] the time constant for removal of H₂O₂ is inversely proportional to [GPx]×[GSH].  This latter phenomenon is the result of the slow recycling of GPx and is consistent with the observations in the animal studies mentioned above.  This correlation between the model and the biological observations confirms that the global biological relationships between [GSH], [GPx], and [H₂O₂] with respect to the ability of cells and tissues to remove H₂O₂.  This work exhibits the importance of mathematical modeling for understanding physiological redox behavior.  Attributed to the large difference in time scales between reaction stages pre- and post-GPx limited, a range of time constants for removal of H₂O₂ can be compiled, from which local intracellular concentrations and in vivo reaction rate constants in the absence of direct measurements can be estimated.  This model paves way for our ongoing quest in understanding cellular oxidative stress by justifying kinetics involved. It also provides preliminary parameters for in depth multiscale computations where intercellular oxidative behavior with related intracellular factors will be simulated. Finally, mathematical modeling of this system can be used to optimize gene therapies in cancer treatment where redox enzyme expression is altered.