Stable isotope labeling is a good way to go.  You could D2O or H218O label the hydroxy metabolite in vivo (see reference).  Or spike a bolus of 13C-labeled parent molecule and track that through to the metabolic by-products.  The latter experiment is more expensive, but gives better results.

Poster Centroiding Error and Isotopic Vectors for Improved LC/MS Analyses

Thank Samuel and Luke for the comment.

I attached a simple graph below to better describe the question.

If XCONHROH is available, a side-by-side metabolism experiment with this hydroxyl metabolite and the parent amide should shed some light on the question.  Also, the signal-time courses for the parent amide and the related metabolites should offer some insights.  I am curious if there are other ways to investigate this kind of problem.

This is what is called a metabolic flux measurement.  Stable isotope labeling of the X is the easiest and most certain way to get the measurement that you want.  You add a small bolus of the *XCONHR to the sample and monitor the isotopic pattern change of the hyroxyide and amine over time by MS.  I am assuming that you can separate these by TLC or LC and then run MS.  Isotopic Vector Angle Analysis is the most accurate way to trace the kinetics.  You might be able to infer something by looking at the R and ROH liberated, assuming that there are no other pathways for R and ROH synthesis .  However, it's not clear whether or not the OH participates in the de-amidation, this would likely depend on where in the R-group the hydroxyl is located.

Thank you, Luke, for the interesting tips.

The abundances and signal-time courses for the cleaved R and ROH would possibly provide valuable understanding of the pathways, as you suggested.  However, it's challenging to detect the R and ROH from this amide by LC-HRMS, the technique we are currently using. 

If you can see all three X species by LC-HRMS, then you could try to infer an answer from an H2[18O] bolus experiment.  The  XONHR[18O]H should increase after labeling and will decrease after the H2[18O] is flushed from the system, if the hydroxide is converted to the amide.  The speed of increase and decrease  of the 18O label on the hydroxide and the relatiive ratio of 18O reached (compared to H2[18O] to normal water in the bolus should be related to the flux of the hydroxide.  If the hydroxide is not involved in the reaction, then the 18O content will be low and the kinetics of rise and fall of 18O will be low.  If it is involved then the rise and fall will be fast and the 18O content should reach that in the water feed.  Its not quantitative, but is cheap and qualitative test.