k⊥ factorization vs. renormalization group: a small-x consistency argument

I investigate, at leading twist level, the consistency of the BFKL equation with the renormalization group, when next-to-leading log-x termss are included in the quark-sea channel of the anomalous dimension matrix. By use of k⊥-factorization, I find that, besides next-to-leading small-x resummation formulae, a leading, x-dependent redefinition of initial quarks and gluons is needed. Its interpretation and phenomenological relevance are briefly discussed.