SoylentNews is people
SoylentNews
from the everything-eventually-kills-you dept.
Arthur T Knackerbracket has found the following story:
It has been accepted wisdom for many years that the more good cholesterol people have in their blood, the better. But the good cholesterol, also known as HDL, might not be as good as we think.
In any case, the results of a new study from the University of Copenhagen seriously contradict the assumption that high levels of HDL in the blood are only a good thing. The researchers have shown that people with extremely high levels of good cholesterol have a higher mortality rate than people with normal levels. For men with extremely high levels, the mortality rate was 106 per cent higher than for the normal group. For women with extremely high levels, the mortality rate was 68 per cent higher.
"These results radically change the way we understand 'good' cholesterol. Doctors like myself have been used to congratulating patients who had a very high level of HDL in their blood. But we should no longer do so, as this study shows a dramatically higher mortality rate," says Børge Nordestgaard, Professor at the Department of Clinical Medicine and one of the authors of the study.
-- submitted from IRC
(Score: 3, Insightful) by PiMuNu on Thursday August 24 2017, @09:08AM (8 children)
From the Fing Journal Article
> with both extreme high and low concentrations [of HDL] being associated with high all-cause mortality risk.
Further
> the multifactorially adjusted hazard ratios for all-cause mortality were 1.36 (95% CI: 1.09–1.70) for men with HDL cholesterol of 2.5–2.99 mmol/L (97–115 mg/dL)
> and 2.06 (1.44–2.95) for men with HDL cholesterol ≥3.0 mmol/L (116 mg/dL)
i.e. there is about a 30 % probability that NULL hypothesis is correct (and there is no result).
(Score: 2) by MostCynical on Thursday August 24 2017, @09:17AM (3 children)
And there is a chance the excessive *anything* will be bad for you.
The difficulty is finding out what is 'excessive' (experimenting *can* be fun, but is often fatal, but then, so is life (so far))
"I guess once you start doubting, there's no end to it." -Batou, Ghost in the Shell: Stand Alone Complex
(Score: 2) by c0lo on Thursday August 24 2017, @11:03AM (2 children)
I can tell, for sure, 100% HDL flowing in your veins and arteries is fatal.
https://www.youtube.com/watch?v=aoFiw2jMy-0 https://soylentnews.org/~MichaelDavidCrawford
(Score: 2) by MostCynical on Thursday August 24 2017, @11:28AM (1 child)
Stop injecting avocado.
"I guess once you start doubting, there's no end to it." -Batou, Ghost in the Shell: Stand Alone Complex
(Score: 2) by Azuma Hazuki on Thursday August 24 2017, @04:09PM
Yeah he'll NEVER be able to afford a house like that!
I am "that girl" your mother warned you about...
(Score: 2, Informative) by shrewdsheep on Thursday August 24 2017, @11:11AM (2 children)
Not sure where you get the 30% from. If a confidence interval does not cover a reference value (1 in this case), there is at most a 5% chance that the data we see was produced under the reference value (very roughly). For the CI (1.44-2.95) the p-value is probably on the order of 10E-3.
Also, the p-value does not translate into a probability that the null hypothesis is correct. Turn it around: if the null hypothesis is true, we have a X% chance (X being the p-value) of observing the data we do or something "more extreme". If X is small (for some definition of small) we consider the combination of null hypothesis and data too unlikely. Let us look for a better hypothesis (and therefore we reject the current null hypothesis).
(Score: 2) by PiMuNu on Thursday August 24 2017, @11:52AM
You are right, I screwed up my statistics.
(Score: 0) by Anonymous Coward on Thursday August 24 2017, @11:54AM
No, this is still transposing the conditional. P(Chance|Data) != P(Data|Chance), just like P(Clouds|Rain) != P(Rain|Clouds). Is the latter true even "very roughly"? No.
You seem to be getting at it with your later sentences but have perhaps not really internalized the consequences of this understanding yet. Also, whose hypothesis was it that measured HDL levels have exactly zero correlation with reported mortality rates? Seems pretty implausible to me...
(Score: 0) by Anonymous Coward on Thursday August 24 2017, @11:52AM
The NULL hypothesis still contradicts "more HDL is better".