Hi Borys,

there are a number of possible side reactions in this case. One is the deprotonation of the ortho position relative to the nitro-group with stabilisation of the the resulting Li-salt by co-ordination with the negatively charged oxygens of the nitro.

Organolithiums can also attack the electrophilic N of the nitro-group directly, in the way Grignards do in the Bartoli Indole synthesis.

I guess, that by fine-tuning the reaction temperature and the solvent you might be able to favour your desired reaction path and improve the yield of the ketone.

It is known that organolithium compounds are able to add to aromatic nito compounds in o-, and p-position. The resulting addition products are relatively stable. Mayby this side reaction causes a not negligible loss of nucleophilic reactand.

The nitro group at para position can activate carbonyl of ester more than other para substituted benzoates. Then, as a parallel reaction organolithiated agent may add twice to carbonyl and 2-(4-nitrophenyl)-1,3-di(pyridin-2-yl)propan-2-ol forms.

All side reactions in the answers of Wiedenau, Allmendinger and Shaabanzadeh are true and possible in your reaction. I think is better change the reaction conditions to obtain better results, for example convert the lithium derivative in the corrisponding Grignard and using it in the presence of a catalytic amount of Fe(III).