Best medical journal (Lancet), proper sample-size (579), from first glance it looks OK!
From their discussion there are some caveats:
- mostly white people checked (so if you're Asian, this might not happen)
- MRI has a known bias towards participation of healthy people, as unhealthy people don't like to hear the "news" about their condition
- the study is "cross-sectional", i.e., the MRI is just from one point in time. Having several measurements from the same people over a period of time (in years) would be much, much better.
P.S.: To lift your mood, the first comment in the news-article is hilarious: "now the government obama care will start to penalize people with high blood pressure? absurd, they just need more cash."
This is consistent with parallel data from lipid research. To wit, some people have low cholesterol due to medication. This reduces their risk of MI by X%, say 40%. Other people have the same level of cholesterol due to genetic effects. From the statin trials, we would think this would correlate with a similar risk reduction. Instead, those with a genetic basis for low cholesterol appear to be protected even more than one would anticipate.
This effect has been attributed to the lifelong exposure to low cholesterol levels. This is not a closed, certain conclusion, but it is supported by the data.
The parallel would be that lifelong exposure to high blood pressure leads to changes over time and produces accumulating harmful effects. The notion is reasonable on its face.
