Some clinicians call for a broad application of "autismness" that scales from "zero autismness" to "classic autism." This would result in a lot of autism diagnoses.
Others, especially researchers, cling to the DSM-III tradition that autism cannot be comorbid with a known physical diagnosis. This was explained to me in this way: if you know the cause, you can research and treat accordingly. Things we do know to be "medical" are moved out of the DSM on a regular basis. This is a logical strategy for categorizing and treating conditions. You wouldn't study Fragile X the same way you might neurotoxin exposure.
I fall into the "if you know it, classify it" camp. If I know the cause of my condition (and I do), then I should receive medical and educational supports accordingly. "Autism" by itself would be too vague. I have a left frontal lobe incursion from a complicated delivery exacerbated by some other issues. In Minnesota, I am diagnosed with HFA ("autism"), but in California my physical injuries and their expressions were the only diagnoses.
I compare this to "cancer." The term by itself, while somewhat helpful, tells a physician little. However, if we clarify the diagnosis of cancer to the specific type, it can be appropriately studied and (hopefully) treated. "Cancer" is too broad to be medically useful, just as "autism" is often too broad to help determine specific research needed and treatment strategies.
For clarification, my specialty is language comprehension and educational strategies. My research was both on the rhetoric of autism diagnostics and how to best address the language development of students with autism; many of us struggle with the "oddities" of language.
Nothing is odder than being "autistic" in Minnesota but not in my native California. It is one reason I do doubt the label's usefulness except in the most general applications.