Causal language and not just in the headline

A Guardian article on a new epidemiologic study looking at a class of blood pressure drugs and dementia started out with a headline that was doubly misleading:

Blood pressure drugs can halve risk of dementia

First of all, this may just be my own reading problem, but when I first saw the word "halve" my brain saw "have" and interpreted it to mean that the drugs increase risk of dementia. But when I slowed down and re-read it, I was not much happier, because I've developed this intellectual/editorial tic, a tendency to notice and then question statements that imply causation. The Guardian lead reiterated the causal language:

Millions of older people who take drugs for high blood pressure or heart problems can more than halve their risk of developing Alzheimer's disease and dementia, according to research.

For the headline and lead to be supportable, there would have to be a clinical trial randomly assigning people (a lot of them) to get ARB drugs, or not. It seemed unlikely, and in fact, it wasn't the case. Here is the original article from BMJ. Sure enough, right in the article's title, it says that it's a prospective cohort study - a very good one, I'm sure, but designed to look at associations, but not causation. And here's the conclusion of the research article:

Angiotensin receptor blockers are associated with a significant reduction in the incidence and progression of Alzheimer’s disease and dementia compared with angiotensin converting enzyme inhibitors or other cardiovascular drugs in a predominantly male population.

Wordy, but - again - in terms of "association" not an active, causal verb phrase like "cuts risk by half."
Regarding expert comments, I thought the article let the head of research of the Alzheimer's Association get a little carried away with the conclusions of the study: "The prospect of using already existing drugs to help in the fight against dementia is attractive." The head of another Alzheimer's research organization did acknowledge the need for trials to confirm the link.