Duchenne muscular dystrophy is a monogenic disease potentially treatable by gene replacement. Use of recombinant adeno-associated virus (AAV) will ultimately require a vascular approach to broadly transduce muscle cells. We tested the impact of preexisting AAV antibodies on microdystrophin expression following vascular delivery to nonhuman primates. Rhesus macaques were treated by isolated limb perfusion using a fluoroscopically guided catheter. In addition to serostatus stratification, the animals were placed into one of the three immune suppression groups: no immune suppression, prednisone, and triple immune suppression (prednisone, tacrolimus, and mycophenolate mofetil). The animals were analyzed for transgene expression at 3 or 6 months. Microdystrophin expression was visualized in AAV, rhesus serotype 74 sero-negative animals (mean: 48.0 ± 20.8%) that was attenuated in sero-positive animals (19.6 ± 18.7%). Immunosuppression did not affect transgene expression. Importantly, removal of AAV binding antibodies by plasmapheresis in AAV sero-positive animals resulted in high-level transduction (60.8 ± 18.0%), which is comparable with that of AAV sero-negative animals (53.7 ± 7.6%), whereas non-pheresed sero-positive animals demonstrated significantly lower transduction levels (10.1 ± 6.0%). These data support the hypothesis that removal of AAV binding antibodies by plasmapheresis permits successful and sustained gene transfer in the presence of preexisting immunity (natural infection) to AAV.