The U(1)_A symmetry of the QCD lagrangian is broken by an anomaly at any temperature.  At infinite temperature it is expected to be effectively restored. We investigate the fate of the anomaly at temperatures just above the chiral phase transition (that happens at temperatures T_c ~ O(\Lambda_QCD). An effective restoration at the chiral transition point is relevant for the symmetry class of the transition, if it is of second order nature.  We found, by using the maximally chiral fermions on a lattice discretisation, that U(1)_A is effectively restored at temperatures that are just above T_c. This excludes some scenarios for the QCD phase diagram. We show that having control on the chiral symmetry is an essential ingredient for the result.