It is common practice in laboratories to create models of degraded articular cartilage in vitro and use these to study the effects of degeneration on cartilage responses to external stimuli such as mechanical loading. However, there are inconsistencies in the reported action of trypsin, and there is no guide on the concentration of trypsin or the time to which a given sample can be treated so that a specific level of proteoglycan depletion is achieved. This paper argues that before any level of confidence can be established in comparative analysis it is necessary to first obtain samples with similar properties. Consequently, we examine the consistency of the outcome of the artificial modification of cartilage relative to the effects of the common enzyme, trypsin, used in the process of in vitro proteoglycan depletion. The results demonstrate that for a given time and enzyme concentration, the action of trypsin on proteoglycans is highly variable and is dependent on the initial distribution and concentration of proteoglycans at different depths, the intrinsic sample depth, the location in the joint space and the medium type, thereby sounding a note of caution to researchers attempting to model a proteoglycan-based degeneration of articular cartilage in their experimental studies.