Here's what happens when you don't control fully for age in an epidemiological study. The first half of the paper is an easy read; the second half needs a quiet room and a clear head.
It is an interesting example of how science works nowadays. Errors are supposed to be picked up by peer review; perhaps blogs do the job better (but you have to be reading the right blogs). Hopefully the authors will publish a correction so that the bibliographic literature contains a reference to the mistake.