At zero temperature and strong applied magnetic fields the ground sate of an anisotropic antiferromagnet is a saturated paramagnet with fully aligned spins. We study the quantum phase transition as the field is reduced below an upper critical Hc2 and the system enters a XY-antiferromagnetic phase. Using a bond operator representation we consider a model spin-1 Heisenberg antiferromagnetic with single-ion anisotropy in hyper-cubic lattices under strong magnetic fields. We show that the transition at Hc2 can be interpreted as a Bose-Einstein condensation (BEC) of magnons. The theoretical results are used to analyze our magnetization versus field data in the organic compound N iCl2-4SC(N H2)2 (DTN) at very low temperatures. This is the ideal BEC system to study this transition since Hc2 is sufficiently low to be reached with static magnetic fields (as opposed to pulsed fields). The scaling of the magnetization as a function of field and temperature close to Hc2 shows excellent agreement with the theoretical predictions. It allows to obtain the quantum critical exponents and confirm the BEC nature of the transition at Hc2. The organic compound N iCl 2 -4SC(N H 2 ) 2 (DTN) undergoes a field induced non-magnetic to XYantiferromagnetic transition 1,2 . This transition can be viewed as a Bose-Einstein condensation of magnons associated with the Ni spin 1 degrees of freedom. Other magnetic systems with a singlet ground state either with spin-1 Ni atoms or spin-1/2 dimers have also been shown to exhibit this transition 3,4,5 . At zero temperature it is driven by the magnetic field H that reduces the Zeeman energy of the S z = 1 state until it becomes degenerate with that of the product state of S z i = 0. At this point, H = H c1 , the antiferromagnetic (AF) interactions give rise to a long range ordered phase. Experimentally, the magnetization M at very low temperatures starts to increase above the critical magnetic field H c1 and eventually saturates above a critical magnetic field H c2 1, 2,6,7,8,9 . The transition at H c1 has been intensively investigated, both theoretically 12,13 and experimentally 1,2,6,7,8,9 , while the one at H c2 is much less studied. The DTN is the ideal BEC system to investigate the latter transition since detailed magnetization curves can be obtained close to the critical field H c2 = 12.3T. In other well known BEC systems, as BaCuSi 2 O 6 10 and T lCuCl 3 11 , the critical fields H c2 are 49T and 83T, respectively, and presently can only be reached using pulsed fields. The excellent quality of the magnetization versus field curves obtained in DTN using standard superconducting coils is essential for the scaling analysis presented here.In this paper we study the transition at H c2 . A theoretical approach is more directly developed, starting from the saturated paramagnetic (PARA) phase. We consider decreasing the external magnetic field at T = 0 to the critical value H c2 where the transverse components of the magnetization condense. A scaling approach for this transition has recently been prop...