I think the sequence is more important than just over 50% or any similar figure. A primer with 4Cs at one end and 4Gs at the other is probably more of a problem than primers with 10C G bases that do not form hairpins .
For PCR normally the GC-content of a primer can be between 45-60%. Besides that it is important not to have repeats (>3bp), especially of GC. At the 3'-end not more than 3 G/C are allowed in the last 5 base pairs. But also a lot is dependent on the desired product, the lenght of the primer, the annealing temperature and the allowed mismatch.