Letter by Noda and Iso Regarding Article, “Low-Density Lipoprotein Cholesterol Concentrations and Death Due to Intraparenchymal Hemorrhage: The Ibaraki Prefectural Health Study”
To the Editor:
We reported an association between low low-density lipoprotein (LDL) cholesterol and increased death due to intraparenchymal hemorrhage in the general Japanese population aged 40 to 79 years in the April 28, 2009, issue of Circulation.1 In his accompanying editorial, Dr Goldstein pointed out that the results were inconsistent between observational studies and randomized clinical trials (RCTs) and concluded, “This suggests no causal relationship between total cholesterol and LDL-C and bleeding risk,” based on the previous RCTs.2 However, the previous RCTs had several limitations and do not necessarily support a conclusion of no causal relationship between low-LDL cholesterol and risk of intraparenchymal hemorrhage.
First, our observational study showed a nonlinear association between LDL cholesterol and risk of intraparenchymal hemorrhage: We observed a negative dose–response association for LDL cholesterol levels <100 mg/dL, whereas the risk plateaued for LDL cholesterol levels >100 mg/dL.1 Therefore, any increased risk during intervention may depend on the LDL cholesterol levels achieved during the follow-up. The RCTs that showed no associations had inadequate statistical power (generally, ≤36%) to detect such associations. Even the Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial, in which participants had relatively low LDL cholesterol, had only 51% statistical power. The post hoc analyses in the SPARCL trial also had low statistical power.
The second limitation is the short intervention period (≈5 years) of these RCTs. A previous meta-analysis of cholesterol-lowering RCTs for coronary heart disease prevention showed that the trials with longer intervention periods had a larger difference between intervention and control groups than did those with shorter periods,3 probably because of the long-term effect of cholesterol lowering. Inasmuch as an effect of low LDL cholesterol on development of intraparenchymal hemorrhage is expected to appear in the long term (ie, over several decades), previous short-term RCTs likely underestimated the association.
Third, the previous RCTs did not discriminate intraparenchymal hemorrhage from subarachnoid hemorrhage. Because the low cholesterol effect is thought to be confined to intraparenchymal hemorrhage,4 which constitutes only 37% of hemorrhagic stroke in Western populations,5 the nondiscrimination of hemorrhagic strokes may have resulted in underestimating the association.1
Finally, the progression of physiological conditions in observational and intervention studies may be different. Observational studies deal with initially low LDL cholesterol levels, whereas intervention studies usually deal with initially high LDL cholesterol, which is reduced to moderate-to-low LDL cholesterol levels during intervention. Furthermore, statins used in the RCTs may have pleiotropic effects (eg, antiinflammatory and antithrombotic effects) rather than cholesterol-lowering effects, which make any causal inference complex.
Whereas RCTs are useful for establishing causal inference, especially where relatively short-term change affects risk, RCTs that are not designed to examine the relationship between low LDL cholesterol and risk of intraparenchymal hemorrhage are not always superior to observational studies. The effect of cholesterol lowering to low levels for a long period has not been tested in previous RCTs, to our knowledge, so only the observational studies are available for this judgment. We need to carefully weigh results from both observational and intervention studies in order to draw causal inferences.
We thank Professors David R. Jacobs and Aaron R. Folsom, School of Public Health, University of Minnesota for their valuable comments.
Noda H, Iso H, Irie F, Sairenchi T, Ohtaka E, Doi M, Izumi Y, Ohta H. Low-density lipoprotein cholesterol concentrations and death due to intraparenchymal hemorrhage: the Ibaraki Prefectural Health Study. Circulation. 2009; 119: 2136–2145.
Goldstein LB. The complex relationship between cholesterol and brain hemorrhage. Circulation. 2009; 119: 2131–2133.
Law MR, Wald NJ, Thompson SG. By how much and how quickly does reduction in serum cholesterol concentration lower risk of ischaemic heart disease? BMJ. 1994; 308: 367–372.
Konishi M, Iso H, Komachi Y, Iida M, Shimamoto T, Jacobs DR Jr, Terao A, Baba S, Sankai T, Ito M. Associations of serum total cholesterol, different types of stroke, and stenosis distribution of cerebral arteries. Stroke. 1993; 24: 954–964.