Galaxy clusters at high redshift are important to test cosmological models and models for the growth of structure. They are difficult to find in wide-angle optical surveys, however, leaving dedicated follow-up of X-ray selected candidates as one promising identification route. We aim to increase the number of galaxy clusters beyond the SDSS-limit, z~0.75. We compiled a list of extended X-ray sources from the 2XMMp catalogue within the footprint of the Sloan Digital Sky Survey. Fields without optical counterpart were selected for further investigation. Deep optical imaging and follow-up spectroscopy were obtained with the Large Binocular Telescope, Arizona (LBT), of those candidates not known to the literature. From initially 19 candidates, selected by visually screening X-ray images of 478 XMM-Newton observations and the corresponding SDSS images, 6 clusters were found in the literature. Imaging data through r,z filters were obtained for the remaining candidates, and 7 were chosen for multi-object (MOS) spectroscopy. Spectroscopic redshifts, optical magnitudes, and X-ray parameters (flux, temperature, and luminosity) are presented for the clusters with spectroscopic redshifts. The distant clusters studied here constitute one additional redshift bin for studies of the L-T relation, which does not seem to evolve from high to low redshifts. The selection method of distant galaxy clusters presented here was highly successful. It is based solely on archival optical (SDSS) and X-ray (XMM-Newton) data. Out of 19 selected candidates, 6 of the 7 candidates selected for spectroscopic follow-up were verified as distant clusters, a further candidate is most likely a group of galaxies at z~1.21. Out of the remaining 12 candidates, 6 were known previously as galaxy clusters, one object is a likely X-ray emission from an AGN radio jet, and for 5 we see no clear evidence for them to be high-redshift galaxy clusters.