Fringing in CST Microstripes

I've recently designed some antennas in CST Microstripes and have built and measured prototypes of these antennas.
In every instance, the prototype antennas have produced a decrease in resonant frequency and an increase in return loss in comparison to the simulations. This implies that for some reason the prototype antennas have a greater total inductance than the simulation models.
My theory is that a possible cause of the increased inductance is that the 'fringing effect' in the prototypes is more than was accounted for in CST, as this would cause the increased inductance.
Is this likely to be the case, with CST not fully accounting for all fringing or is there another explanation for what I have observed?
Many Thanks

It hard to tell without detailed description of the prototypes you've built. It could be the material, feed configuration..etc... If you post more information about your prototypes and some results of what you observed between simulation and measured data, I reckon you will get more response to your query.

No. All EM effects (including fringing) are included in the EM solution.
However, EM simulators are only as good as the user who creates the model. If the meshing is too coarse, you will see differences. Also, the material properties might be different between simulation model and hardware. FR4 is an example of PCB material with large variation in epsilon, depending on the manufacturer.