The distribution of horizontal subspaces is always Lagrangian not only for Riemannian metrics, but also for the Ehresmann connection associated to Finsler metrics and for the Levi-Civita connection of pseudo-Riemannian metrics. It is not always Lagrangian for general sprays though. 

In the Riemannian case this is classical (see Klingenberg's book on Riemannian Geometry or Paternain's book on geodesic flows), but you can also see my answer to this question: https://mathoverflow.net/questions/127319/intuition-for-levi-civita-connection-via-hamiltonian-flows/127394#127394

There is a much more elementary and geometric version of the construction of the connection given in my answer to https://mathoverflow.net/questions/256435/when-is-a-flow-geodesic-and-how-to-construct-the-connection-from-it/256484#256484 . You can complete it by proving (it's really easy!) that the harmonic conjugate---as defined in the answer---of three Lagrangian subspaces is again a Lagrangian subspace.