As usual, correlation is incorrectly assumed to mean causality.

The summary text says there's causality. A few paragraphs into the study, the text interchangeably uses causality and correlation:

> found that carbohydrate intake was associated with increased total mortality.

> By contrast, any type of dietary fat reduced the likelihood of dying.