Small dense low-density lipoprotein cholesterol is strongly associated with NIHSS score and intracranial arterial calcification in acute ischemic stroke subjects

Intracranial artery calcification (IAC) is an important risk factor for cerebral infarction and a key biomarker for intracranial artery stenosis. Small dense low-density lipoprotein cholesterol (sd-LDL-c) was independently associated with increased cardiovascular events and coronary calcification. Our study assessed whether sd-LDL-c is an independent factor for IAC in acute ischemic stroke (AIS) patients. This cross-sectional study involved a total of 754 patients with AIS (mean age: 65 ± 13.2 years). All the patients had received brain computed tomography angiography (CTA) examination to evaluate IAC. Serum sd-LDL-c levels and other biochemical parameters were analyzed. Admission NIHSS score and mRS score at discharge were collected. After 60-days 85 patients died during hospitalization and follow-up. Partial correlation analysis showed that serum sd-LDL-c levels were associated with admission NIHSS score and IAC score after adjusted age and gender. Logistic regression analysis showed that serum sd-LDL-c levels independently predicted NIHSS scores (β = 1.537, 95%CI: 0.134-2.878, p = 0.042) and IAC scores (β = 1.355, 95%CI: 0.319-2.446, p = 0.015). The average level of sd-LDL-c in patients who died was also significantly increased compared to survival patients (1.04 ± 0.59 vs 0.88 ± 0.44 mmol/L, p = 0.017). However, multivariate logistic regression analysis showed serum sd-LDL-c levels could not predict all-cause mortality and prognosis in AIS patients. Our study found that sd-LDL-c as a strong atherogenic lipid particle can independently predict admission NIHSS scores and the severity of cerebral artery calcification in AIS patients. However, its prognostic value in AIS patients still needs further study in the future.

Biochemical measurements. All patients included in the study underwent head CT, electrocardiogram (ECG), MRI, transthoracic echocardiography (TEE) and had standard laboratory examinations at admission. Blood samples were obtained from an antecubital vein and collected in vacutainer tubes containing EDTA. Patients were asked to fast for eight hours before blood draw, and liver enzyme levels, serum lipid profiles including total triglyceride (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-c) and high-density lipoprotein cholesterol (HDL-c), biochemical parameters of the kidney, fasting plasma glucose (FPG), homocysteine (Hcys) and glycosylated hemoglobin (HbA1c) levels were determined on an Olympus AU600 analyzer (Olympus, Tokyo, Japan). sd-LDL-c was measured by the method of Hirano et al. 15 with minor modification 16 using the commercially available assay kit (sd-LDL-c SEIKEN, Denka Seiken Co., Ltd, Tokyo, Japan). Apolipoproteins (apo) A-1, and B were measured by immunoturbidimetry (Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan). The following variables were collected in the acute phase of ischemic stroke: age, gender, cause of ischemic stroke (according to TOAST criteria) 17 , and confirmed or new detected risk factors such as hypertension, diabetes mellitus (DM), coronary heart disease, smoking, long-term alcohol consumption, atrial fibrillation (AF) and BMI. Hyperlipidemia was defined as serum TC was over 5.2 mmol/L or/and serum TG was over 1.7 mmol/L. Those with hyperlipidemia were subgrouped into combined hyperlipidemia, hypertriglyceridemia, and hypercholesterolemia. Combined hyperlipidemia was defined as both serum TC was over 5.2 mmol/L and serum TG was over 1.7 mmol/L; hypertriglyceridemia was defined as serum TG was over 1.7 mmol/L; hypercholesterolemia was defined as serum TC was over 5.2 mmol/L. Assessment of IAC score. All patients underwent multidetector brain CTA using a 64-slice spiral CT device (GE Healthcare, Milwaukee, WI, USA) with the following parameters:120 kVp, 140 mA, 0.9-mm section thickness, 0.9-mm slice acquisition interval, and intravenous administration of 80 mL of iohexol at a rate of 5.0 mL/s. Bone window CT images covered the whole brain (from the skull base to the vertex) to identify IAC. IAC foci were defined as hyperdense foci with a median density greater than 130 Hounsfield units. We used the 5-point scale semi-quantitative scoring system proposed by Babiarz et al. to evaluate IAC 18 . The highest calcification score for each cerebral artery was selected and the calcification score of all evaluated arteries were added to obtain the patient's total calcification score. The severity of IAC was graded based on the total calcification score: 0 was absent, 1-4 was mild, 5-8 was moderate, and 9-12 was severe 19 . Two experienced neurologists independently reviewed the images from CT angiography and graded the degree of cerebral artery calcification.
Statistical analyses. SPSS version 22.0 (SPSS Inc, Chicago, IL, USA) was used for statistical analysis. Continuous variables with a normal distribution were compared using the Student's t-test and ANOVA. Categorical variables were analyzed using the chi-squared test. Patients were divided in tertiles based on their serum sd-LDL-c levels. Partial Spearman correlation coefficients were used to examine the association between serum sd-LDL-c with IAC scores and admission NIHSS scores after adjustment for age and gender. To assess the association between serum sd-LDL-c and severe IAC, binary logistic regression analysis was used. Further analyses were performed to calculate multivariable-adjusted ORs (95% CIs) of severe IAC for the highest tertile versus the lowest tertile of serum sd-LDL-c levels after adjustment for covariates. To determine whether sd-LDL-c and other cholesterol components or ratio predicted short-term all-cause mortality risk of AIS patients, Cox regression model was applied after adjustment for the covariates. Differences were considered significant when P < 0.05.

Result
A total of 754 AIS patients, with an average age of 65 ± 13.2years, including 515 males were enrolled in this study. After 60-days 85 patients died during hospitalization and follow-up. These patients included 632 hypolipemic drugs users and 122 non-hypolipemic drugs users. The average sd-LDL-c level of all patients was 0.90 ± 0.46 mmol/L. There was no difference of sd-LDL-c level between hypolipemic drugs users and non-hypolipemic drugs users (0.89 ± 0.47 vs 1.07 ± 0.37 mmol/L, p = 0.067). We divided non-hypolipemic drugs users into tertiles (0.12-0.64 mmol/L, 0.65-1.07 mmol/L, > 1.07 mmol/L) based on their sd-LDL-c level. Analysis of variance showed that the NIHSS score, IAC score and mortality rate in the highest tertile of sd-LDL-c were significantly higher than in the other two tertiles ( Table 1). The average level of sd-LDL-c in all patients who died was also significantly increased compared to survival patients (1.04 ± 0.59 vs 0.88 ± 0.44 mmol/L, P = 0.017).
We analyzed the correlation between sd-LDL-c level and IAC score and admission NIHSS score in non-hypolipemic drugs users. After adjusting for age and sex, partial correlation analysis showed that sd-LDL-c was correlated with IAC score (r = 0.131, P = 0.006) and admission NIHSS score (r = 0.316, P < 0.001, Fig. 1).
We defined patients with an IAC score above 4 as having severe calcification. The level of sd-LDL-c in patients with severe IAC was significantly increased compared to non-severe IAC patients (1.09 ± 0.55 vs 0.86 ± 0.44 mmol/L, P < 0.001). However, there was no significant difference in admission NIHSS score (6.67 ± 5.78 vs 6.53 ± 6.45, p = 0.873), mortality rate (8.1% vs 12.4%, p = 0.198) between severe intracranial calcification patients and non-severe calcification patients. We assessed the predictive value of TG, TG/HDL-c, sd-LDL-c, LDL-c, ApoB and ApoB/ApoA1 for severe intracranial calcification, admission NIHSS score and mRS score at discharge. After adjusting for age, sex, hypertension, DM, AF, HbA1c variables and so on, logistic regression analysis showed that only sd-LDL-c predicted the severity of IAC (OR = 2.857, 95%CI: 1.549-5.395, p = 0.001, Fig. 2a). However, there was no significant difference in the risk of severe calcification between tertiles groups.

Discussion
In this study, we assessed the relationship between sd-LDL-c levels and IAC, NIHSS scores, short-term all-cause mortality and mRS scores at discharge in AIS patients. The level of sd-LDL-c in patients who died and those with severe IAC were significantly increased. Although there seemed to be a relation between sd-LDL-c and mortality risk and mRS score, sd-LDL-c did not independently predict all-cause mortality and prognosis. sd-LDL-c Figure 1. Partial Spearman correlation coefficients showed the association between serum sd-LDL-c levels with admission NIHSS score and IAC score after adjustment for age and gender. independently predicted the NIHSS score and IAC score in AIS patients, but there was no significant difference in IAC in patients with different levels of sd-LDL-c. As a marker of intracranial atherosclerosis, IAC is an easily measurable and promising biomarker reflecting the severity of intracranial vascular disease. Chen et al. were the first to demonstrate that the incidence of IAC was higher in Chinese patients with ischemic stroke in a cross-sectional study 20 . In a Danish population, Ovesen et al. showed that the severity of intracranial atherosclerosis detected during emergency assessment (graded by the number of cerebral artery calcification) predicts an increased risk of stroke recurrence 21 . The Rotterdam study quantified the volume of IICA calcification measured by CT scan, and determined that IAC was the main risk factor for stroke in white people 9 . Although IAC is often observed, there is currently lack of a standard method to quantify the severity of IAC. Therefore, using severity of IAC in clinical practice is limited.
In our study subjects, sd-LDL-c can independently predict the severity of IAC. sd-LDL-c, as a lipid particle, can easily enter the arterial wall due to its small size. It binds to glycoproteins in the arterial wall and causes oxidative and inflammatory reactions, thereby damaging the intima of blood vessels 22 . In previous studies, we found that sd-LDL-c can independently predict the progression of arteriosclerosis 23 . At the present study, although the cholesterol levels of AIS patients who used lipid-lowering drugs before admission was lower than that of those who did not, there was no difference in sd-LDL-c level between them. The age, IAC score and the proportion of severe IAC were significantly higher in patients who used lipid-lowering drugs than those who did not. It has been suggested that statins may increase the proportion of calcium in coronary atherosclerotic plaque 24 . Simultaneously, a meta-analysis also showed that lowering LDL had no inhibitory effect on coronary artery calcification 25 .We found that LDL could not independently predict the degree of IAC. These information suggest that lowering traditional cholesterol level and inhibiting vascular calcification may be two different pathophysiological mechanisms.
Hyperlipidemia is a well-known risk factor for cardiovascular disease. However, findings of previous observational studies attempting to explain the effects of lipid levels on the prognosis of stroke are variable 13,26,27 . Large artery atherosclerosis is the main subtype of cerebral infarction in the TOAST classification system. However, the proportion of cerebral infarction caused by large artery atherosclerosis in different populations varies 28 . In our study population, the level of sd-LDL-c in patients who had died was significantly increased. Although sd-LDL-c is correlated with mortality risk and mRS score at discharge, it was not an independent predictor of all-cause mortality and short-term prognosis. sd-LDL-c can independently predict the NIHSS score at admission. Zeljkovic et al. conducted a study with a short-term follow-up of 200 patients with new onset of cerebral infarction and found that sd-LDL-c independently predicted the occurrence and short-term mortality of AIS 12 . Our study showed a different result. This may be explained by different subtypes of cerebral infarction in the study population and the length of follow-up.
The NIHSS score has been widely used in evaluating the severity of stroke in patients. TG, LDL-c, HDL-c, TG/ HDL, ApoB and other lipid components or proportions are to some degree correlated with the NIHSS score [29][30][31] . Yasuhiro et al. semi-quantitatively detected sd-LDL-c in peripheral blood of new diagnosed AIS patients 32 . The results showed that sd-LDL-c was associated with adverse prognosis and mRS scores. However, sd-LDL-c was not