Summary: Procedures for deconvolution of pharmacokinetic data are routinely used in the pharmaceutical industry to determine drug release and absorption which is essential in designing optimized drug formulations. Although these procedures are described extensively in the pharmacokinetic literature, they have been studied less from a statistical point of view and variance estimation has not been addressed. We discuss the statistical properties of a numerical procedure for deconvolution. Based on a point-area deconvolution method we define an estimator for the function that describes the time course of drug release from a drug formulation. Asymptotic distributions are derived and several methods of variance and interval estimation are compared