>I think it is somewhat analogous to selling a pure enantiomer once the mixture is running off patent... a theoretical argument can be made for less AEs / better dosing, but i don't think the clinical efficacy is going to improve significantly.<
For a population as a whole, this is probably an accurate assertion for the large majority of cases. For a specific individual, however, removing one isomer from a racemate can make a huge difference in efficacy. (I can attest to this based on first-hand experience.)
Moreover, there are certain cases where a single-isomer “optimization” makes the difference between a blockbuster drug and no drug at all—Exhibit A is Plavix.
>My feeling is that some big pharmas would like to use "optimization" as a means of producing "new" biological drugs, extend patent lives, and prepare to ward off any potential future generic biologic challenges.<
We’re seeing this already.