That's interesting. That strongly suggests that 3-OH-PCM (the N-methyl homologue) would have even more opioid activity. People tend to think that most opioids are based on tertiary amines but tilidine is a prodrug - it's nortilidine that is active (compare strucures. Very similar) and of course dezocine is a primary amine (and thus looks really strange).
I believe that the 3-OH derivative of tilidine is mentioned in a patent BUT I think it was LESS active. I think it requires the ester moiety to be replaced by a ketone (see ketobemidone) because otherwide you end up with 1 enantiomer that is an agonist while the other is an antagonist.
I'm very interested in these small, potent opioids. Tilidine is small (MW 258,349) but has 3 (or possibly 4 is the olefinic bond overlays that of allylprodine or 14-alloxycodeine) moieties (5 have been elucidated) which is why the active enantiomer is as potent as diamorphine weight for weight. It's a forgotten corner of opioid chemistry but we are seeing NMDA/mu affinity overlapping the 2 structures. Camfentamine is also in this class, but lacks the ester function to have SIGNIFICANT opioid activity (it has some).