Cholesterol and Stroke: A Paradox

December 5, 2007 — A new analysis of 61 prospective observational studies has failed to find any association of total cholesterol (TC) with stroke mortality [2]. The research, from the Prospective Studies Collaboration (PSC; Clinical Trial Service Unit, University of Oxford, UK), appears in the December 1, 2007 issue of the Lancet.
Trial coordinator Dr Sarah Lewington (Clinical Trial Service Unit) told heartwire that while the stroke finding "was surprising, it is actually in line with previous observational studies. There has always been this discrepancy between the observational data and the randomized trials." However, she stressed that, despite their findings, "there is conclusive evidence from randomized trials that statins reduce stroke rate to the same degree that they reduce coronary event rates."
She adds that the most important result from the PSC study, to her mind, is that "cholesterol is a risk factor for heart disease not just in middle age, but also in old age. Because the risk in old age is so much greater, the absolute relevance of cholesterol is even greater for older people," she notes. Another key finding was that "high-density lipoprotein cholesterol [HDL-C] adds to the prediction. We found the ratio of TC to HDL is more informative than HDL-C alone and much more informative than TC. So doctors should be measuring HDL-C as well as total cholesterol, if they can. You're much more likely to be able to determine someone's risk if you measure their HDL-C."
No lower threshold for cholesterol
The PSC researchers obtained information from 61 prospective observational studies, mostly in Western Europe or North America, although there were a few studies from China. The total trial population was almost 900,000 adults without previous disease and with baseline measurements of TC and blood pressure. The research was mostly conduced during the 1980s, when TC was routinely measured rather than low-density lipoprotein (LDL) cholesterol, Lewington explained.
During nearly 12 million person-years at risk between the ages of 40 and 89, there were 55 000 vascular deaths (34,000 ischemic heart disease [IHD], 12,000 stroke, 10,000 other). Information about HDL-C was available for 150,000 participants, among whom there were 5000 vascular deaths (3000 IHD, 1000 stroke, 1000 other).
TC was positively associated with IHD mortality in both middle and old age and at all blood-pressure levels. A 1-mmol/L-lower TC was associated with about a half (hazard ratio [HR] 0.44), a third (HR 0.66), and a sixth (HR 0.83) lower IHD mortality in both sexes at ages 40 to 49, 50 to 69, and 70 to 89 years, respectively, throughout the main range of cholesterol in most developed countries, with no apparent threshold.
This latter point is important, says Lewington. "We found no threshold level below which lower cholesterol is not associated with lower risk, so there is no worry about lowering cholesterol as far as heart disease is concerned. Because we had a lot of data we could divide the bottom group up, and even below 3.5 mmol/L there was still an indication that lower cholesterol gave lower risk of heart disease. This shows conclusively there is no risk [to lowering cholesterol]."
The researchers also performed a parallel analyses of the Multiple Risk Factor Intervention Trial (MRFIT) that involved a further 34,242 vascular deaths—the combined results showed similar findings to the PSC study overall.
The PSC researchers found a positive relation between cholesterol and stroke only in middle age and only in those with below-average blood pressure (BP); at older ages (70-89), and particularly for those with systolic BP greater than 145 mm Hg, total cholesterol was negatively related to hemorrhagic and total stroke mortality.
"The absence of any independently positive association between TC and stroke mortality in middle age (after allowing for systolic blood pressure [SBP]) or in those with SBP below 145 mm Hg and the negative association of cholesterol with stroke mortality at older ages or at higher blood pressures are unexplained and invite research," they observe. "Further investigation of exactly how lipoprotein particles affect stroke risks might help to explain this striking discrepancy."
In an accompanying comment [2], Drs Pierre Amarenco (Bichat-Claude Bernard University Hospital, Paris, France) and P Gabriel Steg (Université Paris 7-Denis Diderot, France) note the lack of a clear association between cholesterol and stroke in the PSC study.
But they point out, "The various causes of ischemic stroke might have different associations with cholesterol. While myocardial infarction almost always follows atherothrombotic disease, brain infarction stems from conditions ranging from rheumatic heart disease to atherosclerotic carotid stenosis. Blood cholesterol is associated with carotid stenosis, and carotid stenosis causes stroke, so observational studies including stroke associated with carotid stenosis might mimic the findings with IHD [ischemic heart disease]."
And they note that stroke risk reduction with statins was recently confirmed in Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) in the secondary prevention of stroke or transient ischemic attack: "The lower the achieved LDL cholesterol over the course of the trial, the greater the reduction in the risk of recurrent stroke," they say. Amarenco was in fact the principal investigator of SPARCL, which was published in the New England Journal of Medicine in 2006.
However, they also observe that, as in another trial, baseline LDL cholesterol was not predictive of stroke in SPARCL, and the treatment effect was observed regardless of baseline LDL-C, findings they say are "puzzling."
Nevertheless, "a link between cholesterol and stroke risk probably exists (at least with atherothrombotic stroke), and there is good evidence that lowering blood cholesterol with statins reduces stroke risk and carotid atherosclerosis, independently of blood cholesterol, blood pressure, and age," they state.
Lewington and the PSC group agree: "Treatment should be guided principally by the definitive evidence from randomized trials, that statins substantially reduce not only coronary-event rates but also total stroke rates in patients with a wide range of ages and blood pressures."
Move over LDL: TC/HDL ratio is more informative
With regard to the HDL analyses, the PSC found that the ratio of TC/HDL-C was the strongest predictor of IHD mortality — 40% more informative than non-HDL cholesterol and more than twice as informative as TC.
Amarenco and Steg also comment on this finding: "Interestingly, TC/HDL cholesterol is more informative in this meta-analysis than HDL, non-HDL, or TC. This result parallels the observation in the INTERHEART study that the apolipoprotein B to apolipoprotein A1 (apoB/apoA-1) ratio was the most informative variable.
"These findings argue for applying the benefits of statins to high-risk patients, regardless of age and blood pressure, and suggest that clinicians might need to consider the ratio of TC/HDL cholesterol, rather than the LDL-cholesterol level to which they have become accustomed," the French doctors conclude.
The Clinical Trial Service Unit is involved in clinical trials of cholesterol modification therapy with funding from various companies (Merck, Schering, Solvay) as research grants to (and administered by) Oxford University. One of the study authors works for Oxon Clinical Epidemiology Limited and has stock options in GlaxoSmithKline. Drs. Amarenco and Steg have disclosed various financial relationships with Pfizer, Sanofi-Aventis, AstraZeneca, Daiichi-Sankyo, Merck, Novartis, Boehringer-Ingelheim, Servier, Eli Lilly, Paion, Lundbeck, Eisai, Bristol-Myers Squibb, GlaxoSmithKline, Merck Sharpe and Dohme, Nycomed, Takeda, The Medicines Company, and ZLB-Behring.