I do wonder if there was a confounding factor that they didn't successfully isolate. "General health" seems like an obvious one, but some other factor might be more subtle.

From the abstract: analysis [was] adjusted for age, alcohol consumption, body mass index, systolic blood pressure, smoking status, Type 2 diabetes, previous myocardial infarction, resting heart rate and serum low-density lipoprotein cholesterol [1]. Of course, that doesn't rule out all confounding factors.

[1] http://ageing.oxfordjournals.org/content/early/2016/12/07/ag...

Have you read the study, which is published in a peer-reviewed journal, before making this statement? If not, your comment doesn't add anything useful to the discussion.

Yes, my comment made a statement that added to the conversation. Maybe you missed it? I'll expand for you.

I have studied statistics, and I understand that correlation does not imply causation. However, if you're going to allege that a professional scientific statistician failed to even consider the most trivial and obvious of non-causative correlations, then you're being insulting.

And it's fine to insult scientists. They're not special snowflakes. Do it, if they have a track record of lies, or deceit, or carelessness, or stupidity. But be aware when you're doing it, and why. Don't just blithely say "Hey guys, have we considered the possibility that these scientists are either drooling idiots or evil liars?" That's uncharitable to an extreme.

And yes, I've read it. They are not, as OP suggested, drooling idiots or evil liars.

I misunderstood your comment as agreeing with jey's comment and I concur with your explanation here. My apologies and thank you for expanding.