Not really a problem, relativistic effects are negligible at the very low accelerations MOND is about. Should MOND work with galaxies, a theory that say "if the force is less than a threshold, use MOND, otherwise, use general relativity" is not pretty, but good enough to match the observations, and that's the important part.

There are more elegant ways of doing it (ex: TeVeS), but before looking at the best way to reconcile general relativity and MOND, something we know is possible, it is important to make sure that MOND works at the scales it is supposed to work with. Currently, it doesn't, but dark matter doesn't either. More research is needed, as they say.