Some of this has been mentioned a bit, but I want to make one important point about things like this that could be helpful for others who come across this post later.
First off, yes if you have the same trailing throughout the image, that eliminates coma, tilt, and many other aberrations of the optical system. If the trailing is in the same linear direction (not swirled or radiating out from the center) that eliminates astigmatism. Options that remain; mirror flop/shift, poor polar alignment/PEC/other poor guiding issues, differential flexure, or incorrect guiding parameters. You don't have a cassegrain, so eliminate flop/shift in the mirror (I had this issue on my SCT when guiding with a separate guide scope, the guiding RMS values were good, but the images contained trailing nonetheless). Eliminate (for now) poor polar alignment/PEC/other issues too. Why? Because you said you have good guiding numbers. These issues lead to poor guiding, not trails during good guiding.
You likely have differential flexure (mentioned before) or incorrect guiding parameters. I am almost certain you have the latter as well, and that's because of the point I wanted to make:
The atmosphere and optical setups we have can only guide so well, and the absolute best we could expect to do for composite RMS values, under perfect conditions, doesn't fall much below 0.4" RMS anyways... When people report getting guiding of, say, 0.5" RMS or better, I start to think they might have their PHD2 guiding parameters wrong (I'm referring to focal length, pixel size, and the mount's step size values, etc.) Why? Because if you guided on literally a hot pixel, which shouldn't move at all, you will still get errors of easily 0.3-0.6" RMS, for example. Any guiding better than, say, 0.5" is unlikely to actually be obtained, even in the best seeing with the best mount, with properly set up parameters. It's just not reasonable.
Check these carefully; I once caught an error in my PHD focal length parameter while using a reducer that I didn't initially catch, and I was off substantially. Maybe your actual focal length isn't quite what is advertised. Maybe the camera pixel size is wrong. Etc. Why does this matter? If you find an error here, it will likely increase all of your RMS values, showing the guiding wasn't as good as you thought it was. Then, you can open up some of the possibilities eliminated above, such as poor PA, PE, and the performance of the mount! After all, RMS values are only as useful as the parameters you input, since the latter are what the program uses to convert from arcseconds/px to arcseconds outright.
