Optimal LDL Cholesterol Level for Stroke Patients: Clearing the Cloud to See the Blue Sky:
A Response to Letter by Dr. Castilla-Guerra et al.
Article information
Dear Sir:
No scientific study is without limitations, and the Korean Treat Stroke to Target (TST) study [1] is no exception. As Castilla-Guerra et al. [2] correctly pointed out, the number of patients was relatively small, the follow-up period was short, and the differences in the levels of serum low-density lipoprotein cholesterol (LDL-C) between the low-target group (LT) and high-target group (HT) were less pronounced than those in their French counterparts [3]. However, I believe that the findings of this study are worthy of consideration.
The primary outcome of the TST trial was a composite of two distinct disease events: cardiac (coronary artery disease and urgent coronary artery revascularization) and cerebrovascular (cerebral infarction and cerebral artery revascularization). If the items listed above matter, we can assume that they affect both outcomes similarly. Let’s first look at the side of cardiac outcome; the clinical benefit of LT over the HT was quite evident with statistical significance (hazzard ratio, 0.26 [95% confidence interval 0.09–0.80], P=0.019), which was similar to the French TST data [3]. Considering this aspect, the readers of this paper, including Castilla-Guerra et al. [2], will not consider the abovementioned limitations significant. Why are they inclined to criticize the results of cerebrovascular disease events, even if the number of cerebrovascular events is larger than that of cardiac events?
I personally think that this may reflect the long-term contamination of our mind from the previous arguments, “the lower the LDL-C level, the better the clinical outcome,” that has been repeatedly spoken by influential cardiologists and a few Western neurologists. This argument is consistent with the cardiac outcomes but not with the cerebrovascular outcomes of our study. This problem, as discussed in my previous study [4], seems to persist.
Castilla-Guerra et al. [2] reported that Korean TST data are inconsistent with larger data, such as the Stroke Prevention by Aggressive Reduction in Cholesterol (SPARCL) [5]. However, in my view, both datasets are consistent. First, the benefit of preventing the primary outcome (overall odds ratio) was 16% in the main SPARCL trial. Although significant, the benefit was only half that of patients with internal carotid artery (ICA) disease (33% risk reduction) [6]. Approximately 20 years ago, when SPARCL was presented, this discrepancy was of great concern to me; I thought that the benefit of high-intensity atorvastatin was likely to be much lower (<16% or even nonexistent) in patients with stroke without ICA disease, who probably had etiologies such as intracranial atherosclerosis (ICAS) or small-vessel disease (SVD). This important issue has been neglected and remains a “vague spot,” which has inadvertently been filled with clouds, that is, assumptions and arguments not based on solid scientific data.
The Korean TST study was a small attempt to clarify this issue, at least partially. Here, we compared the efficacy of the LT and HT strategies in patients with different stroke subtypes: SVD, and large artery disease (ICAS or extracranial atherosclerosis [ECAS]). LT was superior to HT in patients with ECAS, whereas the opposite was observed in patients with SVD (Tables 3 and 4 in the previous study [1]). When we looked at only “symptomatic” cases, the P-value for the interaction between the location of cerebral artery disease (ICAS vs. ECAS) and the LDL-C target (low vs. high) was 0.0284. Our data were consistent with those of another Japanese study (Japan Statin Treatment Against Recurrent Stroke [J-STARS]) that attempted to examine this vague spot; pravastatin was effective in preventing recurrent atherosclerotic stroke but not lacunar infarction [7].
Although the efficacy of statins is well understood in patients with atherosclerotic cerebral infarction, a comparison of statin efficacy between patients with ICAS and ECAS has rarely been performed. Difficulties appear to arise from the fact that (1) ICAS is less frequent than ECAS in the Western Hemisphere, (2) its proper diagnosis is costlier than ECAS (requiring magnetic resonance angiography, computed tomography angiography, or conventional angiography vs. simple duplex scan), and (3) it is often difficult to differentiate ICAS from other intracranial diseases, such as moyamoya disease or dissection. Castilla-Guerra et al. [2] emphasized the apparently robust effectiveness of high-dose statins (or low-target LDL-C strategy) in patients with ICAS, mentioning the Stenting vs. Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis (SAMMPRIS) study, in which statins were used to reduce LDL-C levels below 70 mg/dL in the medical arm. Unfortunately, this was not a controlled trial for a lipid-lowering strategy [8], and a subsequent guideline recommended the use of high-dose statins for ICAS based solely on this study [9].
In this Korean TST study, the efficacies of LT and HT were similar in patients with ICAS. However, when we compared “symptomatic ICAS and ECAS,” a clear difference existed in the efficacy of different strategies (LT vs. HT) (Figure 4 in the previous study [1]). As the Korean TST study did not include a control population (those who did not receive statins), it was impossible to identify the appropriate LDL-C targets. I simply wish to say that more studies are needed to determine an optimal LDL-C target in patients with ICAS and SVD, a clouded area due to the scarcity of controlled studies, less careful interpretation of previous study results, and contamination of our minds by loud arguments that are not based on strong scientific data [10].