Can guiding hurt resolution in deep sky images?

In short, no. The supposed "blurring" occurs because you are integrating over longer time than otherwise (although some mount don't do any guiding at all and results suggest there is no differences here). In order to obtain better resolution (other than having a very good seeing) you can integrate over shorter and shorter exposures and cull out the worst results and keep the best. Obviously you take the hit with increased RON.

I am wondering if guiding can sometimes reduce resolution.

Yes, sometimes it can.  There are multiple components that make up any feedback driven control system.  The most obvious component is the amplitude response of the system.  This requires that an image shift of say 5 arc-seconds is met with an equal shift of 5 arc-seconds by the mount to keep the image centered.  The next component of a control system is the frequency response of the system.  The frequency response relates to how quickly the system can respond to a change.  If the seeing is moving the position of the star at a rate of say 1kHz, most mounts will have a hard time accurately tracking any movement at that rate.  Fortunately atmospheric tilt generally changes at a pretty slow rate.  Finally, the least obvious characteristic of a feedback loop is the phase response and this is where it's very possible for tracking to make things worse.  For example: If the time for the signal from the guide camera to be taken and processed takes half the time (or multiple thereof) of the period of a periodic shift in the guide star, the correction signal will be applied out of phase of the image shift and the result will amplify the image motion on the sensor--rather than reducing it.   I've greatly simplified things here.  Phase shifts in the response can be caused by all sorts of things in the system but it is a very important characteristic of any control system.  Phase errors in tracking systems are often called "chasing the seeing", which often happens when the guide rate is set to be too high relative to how fast the system can respond to changes.  The best way to implement a guiding system is to configure it to compensate for small, slowly changing errors in the system (such as PE errors in the drive) and slowly varying changes in the atmosphere.  The idea is to minimize slowly varying errors while ignoring the high frequency stuff that's pretty hard to correct--at least by using the mount.  

BTW, there are devices called, "Active Optics" (in the amateur world), which use a small window to correct image shift at a much higher rate than can be done by the mount.  With the right target and very careful configuration, these things can work to increase the response of a tracking system.

John