There is no reason that an MD-PhD can't become a surgeon.
1. Infrastructure - research is easier, yes, in certain departments (path, med, peds, neuro etc) because of the culture, infrastructure, mentorship etc. The lack of such support structure necessitates a greater effort on part of the trainee/young faculty but the obstacles are not insurrmountable. I think finding the right mentor is key. Someone who will help you clear some hurdles.
2. Lack of topics/funding - this is not a negative factor. It may seem like there is a lack of appropriate topics but that's because research activity tends to be low in surgical fields and EM etc. For EM in particular, at my institution, there are two md-phd faculty doing research in hyperbaric med and ischemic neuronal injury in the dept. The keys here are finding interesting, fundable topics and to create collaborations with experts (as is emphasized in a nice chapter by Dr. Salter of Salter-Harris and CPM fame). In fact, I would argue that this aspect of research as a surgeon is more desirable than for a non-surgeon. Topics are wide open and funding is not as competitive.
3. Time - typically 5-8 years of res/fellow training in both surgical and non-surgical specialties. Compare to cards/electrophys training. If you've committed to academics, you're in for life surgeon or not.
4. Desirabily as a resident - most but not all programs will see the PhD as a plus. Some will drool others will think you're an intellectual snob. Most will fall in between but generally positive. There are a decent number of mdphds in neurosurg and g surg already. Many ent, ortho, uro programs are looking for mdphds. Even some plastics programs are coming around. Don't forget, DO WELL CLINICALLY ("honors," AOA, letters). Don't fret if things aren't perfect..they rarely are. Just do your best and don't blow thing off.
5. Priorities as a student. My take is excel at both and everyone will want you. If you had to choose, surgery first and research second. If you're bad at research, you lose funding. If you're a bad surgeon, you kill people. Also, by and large, the intellectual leaders in surgery that happen to be mdphd (they do exist) are HIGHLY regarded surgeons not mediocre ones. Also residency training is first a clinical program and second an academic one. Excellent clinical performance can overcome okay or not-so-good research. The reverse rarely happens. However, if you're main focus is research far and above surgery, think twice. You want to become a surgeon-scientist not scientist-surgeon. Also, surgery is less about absolute percision than it is about judgement and approach. Train a monkey to do surgery - yeah, if you're running a butcher shop. Train a monkey (or build a machine as people have already done) to do minipreps, perform hybridizations etc. Not a problem.
*What's important is that you throw away your pre-conceptions before entering a program and keep you're eyes wide open. There will be plenty of people in the future giving you advice and things will be different by the time you're applying for residency positions.
Hopefully matching Ortho 2003
University of Pennsylvania