PBN and derivatives such as NXY-059 are synthesized by the condensation of an appropriate derivative of benzaldehyde with NtBHA or some salt (such as the acetate). E.g., US patent 6,479,697, to AstraZeneca, describes making the sodium salt of alpha.(2,4-disulfophenyl)-N-tert-butylnitrone, AKA "NXY-059" this way.
The resulting bond is relatively weak and cleaves spontaneously to NtBHA and the corresponding benzaldehyde, particularly in acidic media. Helpfully, the '697 patent also states: "The free base form of N-tert-butylhydroxylamine is unstable, tending in particular to undergo aerial oxidation. This is evidenced by the formation of blue colours which indicate the presence of the oxidation product, 2-methyl-2-nitrosopropane. The free base of N-tert-butylhydroxylamine cannot therefore easily be stored as such but must be freshly generated immediately before use on each and every occasion that it is required...." I.e., where you have NtBHA, you also get MNP, its parent spintrap.
Doubtless, when prepared and packaged, pharmaceutical NXY-059 is pure. However, this does not exclude subsequent spontaneous hydrolysis in vitro to make NtBHA. Further, as AZ's '693 patent notes above, in air, the NtHBA then oxidizes to the corresponding parent spintrap, 2-methyl-2-nitrosopropane or "MNP". Thus, in addition to hydrolysis itself, there are other potential variables-- E.g., storage under air or an inert gas. Similarly, in view of the high doses of NXY-059 administered in the Saint trials, the actual percentage amounts of NtBHA and MNP required to explain the neuroprotective activity found in the Saint-1 trial are small.