Marion Nestle is doing a casual review of the literature, tallying sponsor-funded research conclusions as favorable or unfavorable. Her results so far: of 142 industry-funded studies, 130 have been favorable to the sponsors, 12 unfavorable. See here.
This is only the latest of a long line of investigations that show, contrary to the insistence of those who seem to believe that all research should be considered equal regardless of the source, that funding source does affect research outcomes–whether because funders will only fund research that they believe will find favorable results (thus skewing the whole of the body of research literature, much like publication bias does) or because funding itself subtly alters any number of procedures, biasing the results (in effect, I believe that this is not much different than the former effect). In both cases, a body of literature is produced that systematically favors sponsors and thereby distorts our picture of reality. See studies here and here.
In order to correct this distortion of reality, which is a systemic fact even if it does not apply in every case, I believe it is both legitimate and necessary to criticize studies on the basis of their funding sources ALONE. It may be impossible to prove distortion in any individual case; but if the sum result of studies points to distortion, does this not cast a shadow on individual cases?