Apolipoprotein B/AI ratio as an independent risk factor for intracranial atherosclerotic stenosis

Participant characteristics

A total of 2210 participants were included in the study, among whom 1138 patients with acute ischemic stroke (359 with ICAS, 779 without ICAS) and 1072 non-stroke controls (239 with ICAS, 833 without ICAS). The apoB/AI ratio was significantly increased among patients with ICAS (mean 0.86±0.27, range 0.25-1.74) compared with patients without ICAS (mean 0.78±0.26, range 0.02-1.78, p<0.001). Baseline characteristics of all the participants are summarized in Table 1. As the results showed, in both stroke and non-stroke groups, age, sex, weight and BMI were not significant when compared patient with ICAS and without ICAS. There is no difference in apoAI level, but apoB level and the apoB/AI ratio differed across the SICAS and SNCAS groups. There are statistic significances in the presence of hypertension and level of glucose between the SICAS and SNCAS groups. In the comparison of the NICAS and NNCAS groups, the apoAI level, apoB level and their ratio showed apparent differences. Other significant factors including presence of hypertension, diabetes, level of glucose, HDL-c and smoking. The tests of interactions between HDL-c, LDL-c and apoB/AI ratio appeared non-significant results (In stroke group, HDL-c p for interaction=0.246, LDL-c p for interaction=0.517; In non-stroke group, HDL-c p for interaction=0.774, LDL-c p for interaction=0.211). The SICAS (0.88±0.26) group had the highest apoB/AI ratio, followed by the SNCAS (0.84±0.26) group, the NICAS (0.82±0.27) group, and lastly the NNCAS (0.73±0.24) group. The discrepancies and the participant distribution among the 4 groups were recorded in Figure 1.

Figure 1. The level of apoB/AI ratio and distribution of participants. Data are represented as mean±SD; ICAS, intracranial atherosclerotic stenosis. SICAS, stroke with ICAS; SNCAS, stroke without ICAS; NICAS, non-stroke with ICAS; NNCAS, non-stroke without ICAS.

Association of apoB/AI ratio and ICAS in stroke group

All covariates were incorporated into the multivariable logistic regression analysis of the SICAS and SNCAS groups. Findings of multivariable analysis to examine the association between apoB/AI ratio and ICAS are provided in Table 2. After adjusting these covariates, the apoB/AI ratio (OR 2.80, 95%CI 1.45-5.42, p =0.002) was demonstrated to be an independent risk factor for ICAS against a background of stroke. Other risk factors included the presence of hypertension (OR 1.52, 95%CI 1.12-2.08, p =0.008) and high level of glucose (OR 1.05, 95%CI 1.01-1.10, p =0.015).

Participants were stratified into quartiles according to the distribution of their serum apoB/AI ratio. (Supplementary Figure 2A showed the distribution of subjects according to quartiles in four groups). For the SICAS and SNCAS groups, the regression analyses indicated that patients with higher apoB/AI ratio quartiles were more inclined to suffer ICAS, compared with the first quartile. Specifically, with the first quartile as the reference category, the ORs and 95%CI were 1.22(0.81-1.82) in the second quartile, 1.71(1.11-2.63) in the third quartile and 2.06(1.27-3.35) in the fourth quartile. The third (p=0.014) and forth (p =0.003) apoB/AI ratio quartiles were significantly associated with SICAS. We also conducted the trend test and final got the p value for trend (p =0.002) which was significant and it also indicated a dose-response relationship between apoB/AI ratio and the risk of suffering ICAS.

Association of apoB/AI ratio and ICAS in non-stroke group

In NICAS and NNCAS groups, the regression analysis showed that the apoB/AI ratio (p <0.001) was still a significant risk factor of suffering ICAS. When referenced to the first quartile, ORs and 95%CI in the second quartile were 1.15(0.71-1.86), in the third quartile were 1.71(1.04-2.82) and in the fourth quartile were 2.00(1.16-3.49). The third (p =0.033) and fourth (p =0.012) quartiles were different between the two groups, which supports the apoB/AI ratio was an independent risk factor for ICAS. The p value for trend (p=0.006) was still remarkable in non-stroke population. Similar to stroke groups, the presence of hypertension (OR 1.91, 95%CI 1.31-2.82, p <0.001) and high level of glucose (OR 1.08, 95%CI 1.01-1.16, p=0.021) were independent risk factors for ICAS in the present study. These associations were still robust after controlling for covariates.

ApoB/AI ratio and extent of steno-occlusion lesions

Table 3 manifests that increased apoB/AI ratio is associated with the extent of steno-occlusion lesions. (Supplementary Figure 2B shows the distribution of subjects according to quartiles in three groups). Comparing the no-ICAS group and the group of 1-3 stenosis arteries, the discrepancy was significant (p <0.001), suggesting that the high level of apoB/AI ratio is a valuable risk factor for atherosclerotic stenosis. However, the comparison between the group of 1-3 stenosis arteries and the group of 4 or more stenosis arteries (p =0.081) failed to demonstrate the association.

Study population

Study subjects were prospectively recruited from January 2014 to June 2018 among patients referred to department of neurology of Qingdao Municipal Hospital for suspected stroke and individuals underwent comprehensive health screening, including brain MRI and magnetic resonance angiography (MRA), at health screening center of Qingdao Municipal Hospital. The non-stroke controls are participants from health screening center who did not have serious health problems as well as TIA or acute ischemic stroke confirmed by neuroimaging. We excluded individuals who: (1) less than 40 years old; (2) underwent incomplete vascular imaging and laboratory tests; (3) had been on statin or fibrate before admission because these drugs could affect the apoB/AI level; (4) had atrial fibrillation, cardiac embolism, vascular disease and had underwent replacement; (5) had intracranial and external artery dissection, arteritis, moyamoya disease, muscular fiber dysplasia; (6) had infection, nausea, tumor, chronic liver disease and renal insufficiency; (7) had history of ischemic stroke. Written informed consent form was obtained from all participants or their legal representatives. This study was approved by the Institutional Ethics Committees of Qingdao Municipal Hospital.

Finally, 1072 non-stroke controls and 1138 stroke patients were included in this study for the analysis (Flowchart of the screening process about the included participants was listed in the Supplementary Figure 1).

Assessment of clinical risk factors

Participants were assessed according to stroke registry information, including demographic profiles, living habits and biomarker data of stroke or vascular risk [24]. Age, sex, weight, body mass index (BMI), history of hypertension, history of diabetes, alcohol consumption, smoking, and high level of glucose were recorded. Before the measurement of blood pressure, subjects were required to rest for at least five min. All subjects were measured at least twice to obtain stable blood pressures. Systolic blood pressure over 140 or diastolic blood pressure over 90 mmHg were regarded as hypertension [25]. Diabetes mellitus was defined as a level of fasting plasma glucose≥7.0mmol/L. Alcohol abuse was defined as moderate to severe alcohol consumption in social history (>168g/week) [24]. We also considered those who smoked regularly at least one cigarette per day at the time of presentation as current smokers [24].

Laboratory measurements

Blood samples of all the participants were drawn after an overnight fast during 24 hours of stroke onset. Serum levels of HDL-c and LDL-c were assayed by enzymatic techniques (Beckman Coulter (AU5800), USA), and the apoB and apoAI levels were assayed by immunonephelometry using a BN II analyzer (Siemens Healthcare, Marburg, Germany) [16]. Some studies have demonstrated that apoB/AI ratio could keep stable within a time interval from stroke onset to the following four weeks [16, 26]. Therefore, the time interval from stroke onset to the measurements of apoB and apoAI affects the validity of the findings scarcely.

Assessment of ICAS

All enrolled participants were evaluated through neuroimaging and vascular imaging, including brain MRI (magnetic resonance imaging) [16] and 3D time-of-flight MRA (magnetic resonance angiography). Brain MRI and MRA were performed with a 3.0 tesla unit. Diffusion-weighted MRI (DWI) was used to define the presence of acute ischemic stroke, and MRA was performed to determine the location and extent of ICAS. Doppler ultrasonography and contrast-enhanced MRA were used to evaluate the extracranial carotid arteries excluding a cardiac or carotid artery source of embolus that might be another etiology of stroke. [2].

The intracranial arteries we evaluated included the intracranial segment of internal carotid and vertebral arteries, basilar artery, the proximal segment of middle cerebral artery (M1, M2), anterior cerebral artery (A1, A2), and posterior cerebral artery (P1, P2). We defined ICAS as atherosclerotic stenosis >50% or the occlusion of the above main intracranial arteries [24, 27]. For the accuracy to prove the association between apoB/AI ratio and ICAS, we exclude the extracranial segment vascular stenosis (Supplementary Figure 3 and Figure 4, showed several representative MRI and ultrasound pictures of ICAS). Two experienced radiologists who were blinded to clinical data interpreted the angiographic features, and disagreements over the presence of stenosis on angiography were solved by consensus.

Subject grouping

All participants were categorized into four subgroups based on an angiographic study, including (1) stroke with artery stenosis (the SICAS group, ischemic stroke with significant atherosclerotic stenosis in the intracranial arteries); (2)stroke without artery stenosis (the SNCAS group, ischemic stroke without significant atherosclerotic stenosis); (3)non-stroke with artery stenosis (the NICAS group, no ischemic stroke but with significant atherosclerotic stenosis); (4)non-stroke without artery stenosis (the NNCAS group, no ischemic stroke without significant atherosclerotic stenosis). To assess the association between apoB/AI ratio and extent of intracranial atherosclerotic steno-occlusion lesions, all participants were further divided into three groups according to number of affected intracranial arteries: no lesion of ICAS (n=1612), 1-3 intracranial stenosis (n=537), 4 or more intracranial stenosis (n=59).

Statistical analysis

All parameters were presented as mean±SD (standard deviation) or proportions appropriately. Wilcoxon test was used to compare intergroup difference for continuous variables, while Chi-squared test was used to examine categorical variables.

We compared baseline variables between stroke and non-stroke groups, SICAS and SNCAS groups, NICAS and NNCAS groups by the above methods. We also evaluated the association between apoB/AI ratio and steno-occlusion lesions among groups using Kruskal- Wallis test. Multivariate logistic regression analyses were performed to explore whether level or quartile of apoB/AI ratio was independently associated with the presence of ICAS, all covariates were included in the model. The lowest quartile was defined as the reference group. Multivariate logistic regression analysis was further performed to measure the relationship between apoB/AI ratio and ischemic stroke. Results are presented as ORs with their 95% CIs. The tests of interactions between HDL-c, LDL-c and apoB/AI ratio were conducted in both stroke and non-stroke groups. We also conducted the power analysis using G*Power 3.1.9.4. The group sample sizes of 359 and 779 in stroke patients achieve 85% power and the group sample sizes of 239 and 833 in non-stroke controls achieve 99% power to detect the difference of apoB/AI between ICAS and NCAS groups. Overall, other statistical tests were conducted using R software version 3.4.2, and the significance level was set at p <0.05 for all data.