Shears... Aslan might be able to answer your questions better, but here are my thoughts...
Where does Fab-anti-VEGF need to be concentrated?... It would be best if it was concentrated where Bruck's membrane is being breached by the new blood vessels... i.e. a sub-retinal implant would be closer to the site of action... That is why they use Fab instead of the whole antibody... The 48 Kd Fab fragment penetrates the anterior limiting membrane of the retina, whereas the 150 kd Ab does not (Gaudreault et al, 2005, IOVS 46:726).
How much VEGF is actually floating around inside the vitreous?.... Aqueous humor concentration of VEGF in AMD is about 500 pg/ml (Tong JP et al., Am J Ophthalmol 141:456, 2006); concentrations are about 2500 pg/ml in the vitreous of diabetic retinopathy (Ogata et al. Am J Ophthalmol 134:348, 2002).
Does the vitreous VEGF need to be overwhelmed by an injection?... When the Ab is injected as a solution all of the molecules are available to bind to VEGF.. An injection of 1 mg or so, will not be saturated during the first month... But, I am suggesting that when the Ab is on an implant, then only the surface molecules are in play, so VEGF binding to the implant would be like rust accumulating on a nail... It would bind and, perhaps, saturate the surface Ab molecules; and, it would be these molecules on the surface which are next in line to be released.
I caution that these musing might be circumvented by some clever surface science that I don't know about.