查到了最后引用的一片文章，结论挺有意思。

文章选用的指标胆固醇是脑卒中后72小时的值，然后长期追踪。并不是长期保持一个高胆固醇，存活率更高。


J Stroke Cerebrovasc Dis. 2014 Jan;23(1):e47-53. doi: 10.1016/j.jstrokecerebrovasdis.2013.08.009. Epub 2013 Oct 6.



High cholesterol levels are associated with improved long-term survival after acute ischemic stroke.

Markaki I1, Nilsson U2, Kostulas K2, Sj?strand C2.
Author information
Abstract

BACKGROUND:

Prior statin treatment and high admission cholesterol have been associated with favorable outcome after ischemic stroke (IS), a paradox not completely explained. The aim of this study was to investigate the effect of admission cholesterol levels and the impact of statin treatment on short- and long-term survival after IS.
METHODS:

Consecutive patients admitted in 2006 and 2010 were included in the study. Total cholesterol of 4.6 mmol/L or more was defined as high. Logistic regression analysis was performed to assess predictors of 1-month mortality, and Cox proportional hazard regression analysis was applied to investigate predictors of long-term mortality.
RESULTS:

Of 190 patients included in the final analysis, 21 (11%) died within 1 month and 61 (32%) died during 7 years of observation. Low cholesterol was associated with older age, lower blood pressure (BP), presence of angina, and higher risk of death. Three-month, 1-year, and 5-year survival rates were 100%, 98%, and 84%, respectively, in high cholesterol patients, compared with 92%, 87%, and 57% in low cholesterol group (P = .0001 with the log-rank test). Mortality risk was increased for patients with low cholesterol (hazard ratio: 1.97; 95% confidence interval [CI]: 1.05-3.69), after adjustment for age and admission National Institutes of Health Stroke Scale score. After further adjustment for angina and admission BP, the effect of cholesterol on mortality risk was still obvious, yet attenuated (hazard ratio: 1.87; 95% CI: .94-3.32).
CONCLUSIONS:

High admission cholesterol may be associated with increased long-term survival after IS. Future studies on the temporal profile of cholesterol levels and stroke outcome would be of interest.

所有跟帖： 

• 样本太小。而且，选择的人群特殊，影响因素多样。 -相信科学- ♂ (350 bytes) () 10/17/2014 postreply 13:47:37

• 文章作者也提到了你所说的可能偏差。 -xuxinxuexi888- ♂ (0 bytes) () 10/17/2014 postreply 13:54:51

• 这种为了发表而发表的文章一钱不值。养生月刊？拉倒吧。 -相信科学- ♂ (0 bytes) () 10/17/2014 postreply 13:56:59

• 你提出的是对的, 也只有初发病时的数值, 才有意义 -益生菌- ♂ (0 bytes) () 10/17/2014 postreply 16:26:34

• 喜欢这种认真。 -欲千北- ♂ (0 bytes) () 10/17/2014 postreply 20:09:11

• 有原文吗？ -欲千北- ♂ (0 bytes) () 10/17/2014 postreply 21:04:53