Post hoc analysis of a randomized, double-blind, prospective trial evaluating a CXCR1/2 inhibitor in new-onset type 1 diabetes: endo-metabolic features at baseline identify a subgroup of responders