This is not idle unproven speculation. Scientists use phrases like "suggests" rather than "proves" only because they recognize their own data can be misleading. This is published in one of the most competitive journals, speculation doesn't cut it.
The article can be found paywalled here
The abstract highlights that any uncertainty is in the related details, not whether or not it happened.
Clostridium difficile disease has recently increased to become a dominant nosocomial pathogen in North America and Europe, although little is known about what has driven this emergence. Here we show that two epidemic ribotypes (RT027 and RT078) have acquired unique mechanisms to metabolize low concentrations of the disaccharide trehalose. RT027 strains contain a single point mutation in the trehalose repressor that increases the sensitivity of this ribotype to trehalose by more than 500-fold. Furthermore, dietary trehalose increases the virulence of a RT027 strain in a mouse model of infection. RT078 strains acquired a cluster of four genes involved in trehalose metabolism, including a PTS permease that is both necessary and sufficient for growth on low concentrations of trehalose. We propose that the implementation of trehalose as a food additive into the human diet, shortly before the emergence of these two epidemic lineages, helped select for their emergence and contributed to hypervirulence.
I haven't read the paper and don't have a background in it. Reviewers do sometimes make mistakes obviously. But you'd be an idiot to say this is "just an unproven possibility." Leave spewing "meh, scientists, what do they know, just a theory" FUD to the sleazeballs hired by the relevant industry. If you have an actual critique of their methods, by all means, post it here and on pubmed commons or wherever else. Publish a response in nature even. But don't fucking parrot cigarette company lawyers, climate change deniers, and creationists, here on slashdot.