Given their charge and mass, as soon as you start accelerating particles around a loop with a given magnetic field to deflect them, only particles with the correct mass/charge ratio survive. In effect you have built a giant mass spectrometer - other isotopes of hydrogen are too heavy, and the Lorentz forces are insufficient to deflect them down the tunnel.
As was pointed out by @DMcKee, the process of extracting the protons includes bends in the injector - any particles with the wrong Q/m ratio will be eliminated there, before making it into the main accelerator. You can see that in this diagram (from https://lhc-machine-outreach.web.cern.ch/lhc-machine-outreach/images/complex/Cern-complex.gif)
An of course the first part of the acceleration happens in a LINAC. Although it is a straight line, it selects for the right particles as the RF frequency (and the spacing of the acceleration stages) is tuned for a specific Q/m ratio. So anything that is not a proton will almost certainly never make it to the first bend - anything that did, would "skid out" at that point.