The engines-on-the-side configuration is a way to try to deal with engine weight changes. Huge trim issues arise if your engine isn't as light (or heavy) as you think it will be.
Spaceplanes with engines at the back face a real struggle with balance if _anything_ changes in the engines - they tend to be very heavy compared to anything else in the 'dry' structure of the vehicle and a small error either way can leave you with depleted uranium bulkheads to pull your CG back if it's too late in the development programme to change the configuration.
You really can't get away with stuff like that since single-stage-to-orbit is _really_ hard to do.
Couple that with the big aerodynamic centre changes that you tend to get over such a large flight regime and you may end up with a lot of mass budget being taken up by control surfaces.
I understood from a couple of lecturers at Uni and past co-workers that this would have been one of the really big problems that HOTOL would have had to deal with had it not been cancelled... (which project Alan Bond of Reaction Engines was also involved in)