Very interesting. I had the same thought as WisNeuro - wondering about potential moderating effects of that comparison. Actually had limited experience treating PTSD despite my best efforts to get it so have only done CPT (and even that only a handful of cases), but my experience there was definitely that people vary tremendously in their "psychological mindedness" and the value of certain components seems to depend more heavily on that than I imagine would be the case for PE. I could also see differences as a function of the nature of the index trauma. Some experiences are inherently more "cognitive" than others. Even the broad categories (i.e. military combat, sexual violence) have tremendous within-category variation.
Is exposure itself the only active ingredient in both? I'm reminded of the behavioral activation/ vs general CBT for depression literature. In general, BA seems to be the primary driver of outcomes. The literature disagrees with me, but I'm still not convinced that is necessarily reality. There are many potential moderators we haven't explored to the degree that would satisfy me and cognitive constructs in general are tougher to measure.
Clinically, myself and virtually everyone I know who has done depression treatment has encountered patients who do better with more cognitive work than a pure BA approach. Its complicated though, because the patients who "want" a more cognitive approach are also not necessarily the ones who do better with it.