Risk factors for ischemic stroke in China: a case-control study

Introduction

Stroke is one of the primary causes of death and disability in adults worldwide.¹ Ischemic stroke (IS) has brought a huge burden on the families and lives of patients, which also increases economic pressure on patients.² Ischemic stroke accounts for the largest proportion of stroke patients.³ During the last few decades, the incidence of ischemic stroke has gradually increased, particularly in developing countries.⁴ Globally there were 12.2 million stroke cases, 101 million prevalent cases for stroke, 6.55 death cases due to stroke in 2019, which was increased by 70.0%, 85.0% and 43.0% compared with the year of 1990, respectively.⁵ In China, the incidence of ischemic stroke has increased in the past. There were 3.94 million stroke cases, 28.76 prevalent cases and 2.19 million death cases because of stroke in 2019; the incidence for stroke was increased by 86.0%, the rate of stroke prevalence was increased by 106·0%, the mortality was increased by 32.3% compared to the year of 1990.⁶

Some studies have reported biochemical indicators that may be correlated with the occurrence of ischemic stroke. Previous studies have reported an association between uric acid,⁷ homocysteine,⁸ total bilirubin,⁹ and the incidence of ischemic stroke; however, very few studies have indicated that platelet large cell ratio (P-LCR),¹⁰ mean platelet volume (MPV),¹¹ red blood cell volume distribution width (RDW),¹² and serum creatinine¹³ may be correlated with ischemic stroke incidence. Some studies have also suggested that CCA-IMT may affect the ischemic stroke incidence.¹⁴ Because of the lack of effective methods for treating stroke, the most important measure is prevention. Thus, it is crucial to strengthen the studies relating to the risk factors of ischemic stroke and to control these risk factors in order to decrease the incidence and death rate of ischemic stroke. Our study aimed to determine whether uric acid, total bilirubin, serum creatinine, homocysteine, P-LCR, MPV, RDW, and CCA-IMT were associated with the incidence of ischemic stroke and whether there were sex differences in their association.

Methods

The process of screening patients

Using the Patient Records as our Subject Database from the Selected Hospital in China and after obtaining Ethics Approval from the Hospital Ethics Committee: Manually select Study Subjects from the Patient Records using the Number Codes, Method of Selection of Subjects, Method of Randomization, Sample Size, Inclusion and Exclusion Criteria, and Final Samples for Cases and Controls (Figure 1).

Figure 1. Flow chart of the process for screening patients.

Results

The characteristics of all variables are shown in Table 1 and the characteristics of the quantitative variables are shown in Table 2. There was a difference in the levels of UA between case group 318.00(256.00, 383.75) μmol/L and control group 289.50(243.25, 334.75) μmol/L, there was a difference in the levels of Hcy between case group 12.84(11.64, 13.29) μmol/L and control group 11.74(8.58, 12.84) μmol/L, there was a difference in Scr levels between case group 69.00(58.00, 82.60) μmol/L and control group 61.00(53.00, 76.00) μmol/L, there was a difference in TBIL levels between case group 11.40(8.70, 15.78) μmol/L and control group 11.50(9.00, 14.60) μmol/L, there was a difference in FBS levels between case group 6.00(4.80, 6.80) mmol/L and control group 5.20(4.50, 6.08) mmol/L, there was a difference in HDL-C levels between case group 1.18(1.00, 1.30) mmol/L and control group 1.20(1.00, 1.40) mmol/L, there was a difference in SBP levels between case group 145.00(130.00, 160.00) mmHg and control group 130.00(120.00, 142.00) mmHg, there was a difference in DBP levels in case group 82.00(77.00, 90.00) mmHg and control group 80.00(73.00, 87.00) mmHg, there was a difference in the right CCA-IMT between case group 0.90(0.89, 0.90) mm and control group 0.89(0.80, 0.90) mm, there was a difference in the left CCA-IMT between case group 0.90(0.89, 0.90) mm and control group 0.89(0.80, 0.90) mm, there was a difference in age between case group 65.00±11.00 years old and control group 61.00±12.00 years old (Table 2). Among these quantitative variables (p<0.05), UA, Hcy, Scr, FBS, TG, SBP, DBP, right CCA-IMT, left CCA-IMT, and age were higher in cases than in controls, except for HDL-C level. There were no differences in TBIL, MPV, P-LCR, RDW-SD, RDW-CV, TC, and LDL-C levels between the case and control groups.

The characteristics of the qualitative variables are shown in Table 3. There was a difference in sex between cases and controls (p<0.001). In the case group, there were 244 (61.0%) males and 156 (39.0%) females. In total, 77.0% of males had ischemic stroke compared to 23.0% of males without ischemic stroke. There was a significant difference in hypertension between cases and controls (p<0.001). Among the case group, 225 (56.25%) ischemic stroke patients had hypertension compared with 175(43.75.0%) ischemic stroke patients without hypertension. Among all hypertensive patients, hypertensive patients with ischemic stroke accounted for 82.1%, compared to only 17.9% of hypertensive patients without ischemic stroke. There was a difference in diabetes among cases and controls (p<0.001); diabetes patients with ischemic stroke comprised 85.6%, which was more than diabetes patients without ischemic stroke comprising only 14.4%. There was no difference in dyslipidemia and CHD between the cases and controls.

Univariate binary logistic regression results indicated that UA, Hcy, Scr, TBIL, FBS, HDL-C, SBP, DBP, right CCA-IMT, left CCA-IMT, age, sex, hypertension, and diabetes were significant (Table 4) (p<0.05). Variables with a p-value was under 0.05 were included in the multivariate binary logistic regression analysis. There was no difference in MPV, P-LCR, RDW-CV, RDW-SD, TC, TG, LDL-C, CHD, or dyslipidemia in the univariate analysis. In multivariate binary logistic analysis, after adjustment for age and sex, Hcy, FBS, right CCA-IMT, and hypertension were associated with the incidence of ischemic stroke (Table 5, 6) (p<0.001). The odds of ischemic stroke increased with Hcy (adjusted OR=1.217, 95%CI=1.129-1.311, p<0.001), FBS (adjusted OR=1.313, 95%CI=1.169-1.475, P<0.001), and right CCA-IMT (adjusted OR=45.273, 95%CI=6.693-306.232, p<0.001). The hypertensive persons had a higher possibility of ischemic stroke than persons without hypertension (adjusted OR=3.161, 95%CI=2.086-4.790, P<0.001). Similarly, the males had higher odds of ischemic stroke compared to the females (adjusted OR=2.120, 95%CI=1.420-3.164, p<0.001). For every unit increase in Hcy and FBS levels, the possibility of having ischemic stroke increased by 1.217 and 1.313, respectively. For every unit increase in the thickness of the right common carotid artery intima media, the odds of having ischemic stroke increased by 45.273, and the odds of a hypertensive person having ischemic stroke was 3.161 times higher than that of a person who was not hypertensive. The odds of male patients with ischemic stroke were 2.120 times higher than those of female patients. The four factors viz Hcy, FBS, right CCA-IMT and hypertension were independent risk factors for ischemic stroke (p<0.001). We checked the collinearity diagnostics, suggesting that there was no multicollinearity among these risk factors.

According to the univariate and multivariate binary logistic regression results, Hcy, FBS, right CCA-IMT, hypertension, and sex were included in the final model, and the ROC curve was drawn (Figure 2). The AUC value for this model was 0.802 (Table 7), suggesting that prediction model had good discrimination (Table 8).

Figure 2. ROC curve.

The characteristics of the quantitative variables in the male and female groups are shown in Table 9. There was a difference in the levels of UA, Hcy, Scr, TBIL, MPV, P-LCR, TC, LDL-C, HDL-C, SBP, DBP, and right and left CCA-IMT between both sexes (p<0.05). The RDW-SD, RDW-CV, FBS, TG, and age levels did not differ significantly between the male and female groups. After controlling for confounding factors of age, a sex difference was found in MPV with ischemic stroke (Table 10); increased MPV levels were correlated with the occurrence of ischemic stroke in females (adjusted OR=1.309, 95%CI=1.043-1.641, p=0.020) but not in males (adjusted OR=0.987, 95%CI=0.785-1.241, p=0.911). There was no sex difference in the relationship between Hcy, Scr, TBIL, P-LCR, RDW-CV, RDW-SD, and the incidence of ischemic stroke.

Discussion

Previous studies have reported that MPV, P-LCR, and RDW levels are associated with ischemic stroke. However, MPV, P-LCR, and RDW were not associated with ischemic stroke occurrence in this study. Research on the levels of P-LCR, RDW, and the occurrence of ischemic stroke is still scarce. In our study, MPV was not a risk factor for ischemic stroke; this result was similar to that of a few studies, such as the study by Lok et al., who reported that MPV was not associated with ischemic stroke and was not a reliable marker for predicting the prognosis of ischemic stroke.¹⁵ Our results were in agreement with some studies, such as the study by Ha et al., which showed that a higher MPV was associated with ischemic stroke.¹⁶ While the participants in their study were all patients with ischemic stroke and atrial fibrillation, this was not the case in our study. Atrial fibrillation may be an influential factor for MPV levels in ischemic stroke patients, which may affect the results. The point of the study by Özkan et al held the same view as Ha et al’s study, while they selected patients of ischemic stroke patients and also had sinus rhythm in their study.¹⁷ In fact, almost no studies have been dedicated to the levels of MPV and ischemic stroke occurrence. These studies on MPV and ischemic stroke occurrence included participants who were often combined with other specific diseases, but not simple ischemic stroke patients. In addition, most studies have mainly focused on MPV levels and the outcome of ischemic stroke¹⁸ using mRS or NIHSS scores,¹⁹ but have not focused on the relationship between MPV levels and the incidence of ischemic stroke. This may be the reason for the different MPV results. There are very few studies on P-LCR levels and ischemic stroke, and almost no research has been dedicated to the relationship between the levels of P-LCR and the occurrence of ischemic stroke, except for one study by Cui et al. published in 2020 that indicated that P-LCR, MPV, RDW-CV, and RDW-SD were highly correlated with ischemic stroke occurrence.²⁰ This result is in agreement with our results, which indicated that the levels of P-LCR, MPV, RDW-CV, and RDW-SD were not associated with the occurrence of ischemic stroke. This study, with a small sample size, had 41 participants in the case group and 80 in the control group, but without sample size calculation and adequate power, the results of this study could not strongly prove that they had an association with ischemic stroke. There are some studies related to RDW levels and cardiovascular diseases, but few studies have focused on RDW levels in ischemic stroke. Very few studies have held different views about the relationship between RDW and ischemic stroke. Our study showed that RDW was correlated with ischemic stroke occurrence, which was similar to the findings of by Ntaios et al. that RDW cannot predict the severity and outcome of ischemic stroke.²¹ This was in contrast to another meta-analysis published in 2017, which indicated that RDW is a predictor of the incidence of ischemic stroke.¹² The representative references in this meta-analysis had some limitations when studying RDW and the occurrence of ischemic stroke. In these studies, which were selected by this meta-analysis, the participants were chosen from the general population rather than from ischemic stroke patients. This may have affected the results. They concluded that RDW levels were more likely to be associated with total stroke rather than with ischemic stroke²²; some studies were not adequately powered because of the small sample size and lack of sample size calculation,²³ or some used RDW levels to predict the outcome of ischemic stroke using mRS scores. Another study compared RDW levels across ischemic stroke subtypes.²⁴ They may not truly reflect the association between RDW levels and the incidence of ischemic stroke; therefore, they did not provide powerful evidence supporting the correlation between RDW levels and the incidence of ischemic stroke.

In the last few decades, many researchers have investigated the relationship between the levels of ischemic stroke and uric acid, and most of these studies have suggested that the levels of uric acid are related to the incidence of ischemic stroke. In animal experiments, they consistently confirmed that uric acid was associated with ischemic stroke and that uric acid played a protective role in ischemic stroke.²⁵ High uric acid levels or the addition of uric acid were beneficial for the recovery of neurological function and reduced the severity of ischemic stroke. In human clinical data collection research, researchers have held contrasting views on the specific relationship between uric acid and the incidence of ischemic stroke. Some researchers considered uric acid to be a risk factor for ischemic stroke and elevated uric acid levels were associated with the incidence of ischemic stroke²⁶; some considered uric acid to be a protective factor and high levels indicated better outcome for ischemic stroke²⁷ and low levels were correlated with an elevated incidence of ischemic stroke. Others considered that too low or too high levels of uric acid both had a higher occurrence of ischemic stroke than moderate levels, which had the lowest incidence of ischemic stroke.²⁸ The association between ischemic stroke and uric acid was found to be positive, negative, U-shaped, or J-shaped, and there were also sex differences between them. However, studies reporting uric acid as a risk factor were more numerous than those reporting it as a protective factor. Our study showed that high uric acid levels were associated with the incidence of ischemic stroke in univariate binary logistic regression. This result conforms to those of most studies on uric acid and ischemic stroke. After adjusting for the effect of sex in the multivariate binary logistic regression, this significance disappeared, and the uric acid level was not significant. This result was the same as that in the study by Iranmanesh et al., who found no relationship between uric acid and ischemic stroke.²⁹ Another prospective study did not observe that uric acid level was correlated with the incidence of ischemic stroke.³⁰ Many other studies have failed to establish a relationship between uric acid and ischemic stroke incidence after controlling for confounders, such as sex, diabetes, hypertension, or dyslipidemia.³¹ However, in the sex difference for uric acid in ischemic stroke, our study showed that the levels of uric acid were associated with the occurrence of ischemic stroke among females but not males before controlling for confounding factors of age, and after controlling for age, and there was no sex difference in both sexes; uric acid was not associated with ischemic stroke among males and females, which was different from the findings of Khalil et al., which indicated that their association was only in females.³² Our study also showed that Scr levels were associated with ischemic stroke in univariate binary logistic regression, while it was not significant in multivariate analysis; Scr was not a risk factor for ischemic stroke.

Although our study did not find a relationship between uric acid level and the incidence of ischemic stroke after controlling for the confounders of age and sex, homocysteine level was still significant in the final model. Our results indicated that increased homocysteine levels were correlated with an elevated incidence of ischemic stroke, which is in agreement with most previous relevant studies on homocysteine in ischemic stroke.^33,34 In recent years, more researchers have focused on the relationship between the levels of homocysteine and the incidence of ischemic stroke, and they consistently considered homocysteine to be associated with ischemic stroke, but their results contradicted each other in some aspects such as sex differences in ischemic stroke. Zhong et al. found that among persons with high homocysteine levels, females had a stronger connection to the increased incidence of ischemic stroke than males.³⁵ However, Shi et al. pointed out that this correlation was stronger in males.³⁶ Our study inferred homocysteine to be an independent risk factor for ischemic stroke, and that the incidence of ischemic stroke was increased by the increase in homocysteine levels, every unit increase in homocysteine levels, and the possibility of ischemic stroke increased by 1.217. While we did not find any sex difference, homocysteine levels in both females and males were associated with the occurrence of ischemic stroke in our study.

In our study, TBIL levels were not associated with the incidence of ischemic stroke. But according to previous studies^9,37 the levels of TBIL which were used to predict the occurrence of ischemic stroke was somewhat controversial, researchers held different views. Some researchers pointed out that TBIL did not correlate with the occurrence of ischemic stroke,³⁸ while other studies indicated that TBIL levels had an inverse relationship with the occurrence of ischemic stroke, and higher levels of TBIL suggest a lower incidence of ischemic stroke.^39,40 More studies have focused on the relationship between the severity of ischemic stroke and TBIL levels, and consistently found that bilirubin levels had a positive correlation with the severity of ischemic stroke,⁴¹ but few studies have examined the relationship between the levels and occurrence of ischemic stroke. In our study, although TBIL was associated with ischemic stroke in the univariate binary logistic regression, TBIL was not significant in the multivariate analysis. In a limited number of related studies on TBIL and ischemic stroke, more studies proved that TBIL levels had an inverse relationship with the occurrence of ischemic stroke, but we found that TBIL was not a stable and reliable predictor of the occurrence of ischemic stroke. Some acute ischemic stroke patients may have high TBIL levels, while others may have very low levels of TBIL, which did not show a general trend. The results for the association between TBIL and ischemic stroke were more influenced by the study design, selected sample participants, and sample size. To obtain more reliable results for TBIL and ischemic stroke, better designed and more rigorous studies as well as sufficient sample sizes are required. In the biliary markers of ischemic stroke, such as cholyglycine (CG), TBIL, indirect bilirubin (IDBIL), and direct bilirubin (DBIL), CG is a more sensitive and reliable indicator for predicting the incidence of ischemic stroke according to our previous research on biliary markers of ischemic stroke,⁴² but TBIL was not very stable and sensitive. Although some studies reported that TBIL was associated with the occurrence of ischemic stroke, TBIL was not an ideal and reliable predictor of ischemic stroke in these biliary markers. While there are very few studies on CG, researchers could pay more attention to this new predictor for studies on the relationship between CG levels and the incidence of ischemic stroke.

Our results showed that FBS was an independent risk factor for ischemic stroke. Increased FBS levels were associated with an increased incidence of ischemic stroke, which is consistent with most previous studies on blood sugar and ischemic stroke. High levels of FBS can lead to insulin resistance and cause endothelial dysfunction, which damages to the cerebral vasculature, and increases the risk of ischemic stroke.⁴³ For every unit increase in FBS levels, the possibility of having ischemic stroke increased 1.313.

Our results showed hypertension had significant association with the incidence of ischemic stroke. Hypertension was an independent risk factor for ischemic stroke even after adjusting confounding factors in multivariate analysis. Most previous studies showed hypertension had association with the incidence of ischemic stroke.⁴⁴ Our results were consistent with most previous related studies.⁴⁵ Bassa et al. (2022) pointed out that hypertension was associated with the incidence of ischemic stroke and the prevalence of hypertension in patients with ischemic stroke was significantly higher than patients without.⁴⁶ He et al. (2021) also indicated hypertension had association with the incidence of ischemic stroke, in the early acute ischemic stroke stage, maintaining suitable blood pressure levels and small blood pressure variability was beneficial to reduce the recurrence of ischemic stroke.⁴⁷ Hypertension had significant impact on the incidence of ischemic stroke. Hypertension can cause ischemic degeneration of cerebral blood vessels, promote the occurrence and development of cerebral atherosclerosis, thicken the wall for blood vessels of brain, narrow lumen or have plaque rupture secondary to brain thrombosis, and thrombus detachment could cause brain artery embolism, leading to blood supply of brain insufficiency or cerebral ischemia.⁴⁸ High blood pressure may accelerate brain aging and damage neurons. Mitochondria are energy production centers within cells, and it can cause mitochondrial dysfunction, resulting in damage to the nervous system. High blood pressure can promote oxidative stress response, cause endothelial cell damage, cause endothelial cell dysfunction, promote atherosclerosis and thrombosis, and increase the incidence of ischemic stroke. Mitochondria are energy production centers within cells, and high blood pressure also can cause mitochondrial dysfunction, resulting in damage to the nervous system and aggravate ischemic stroke.⁴⁹ Our results indicated hypertension were significantly associated with the incidence of ischemic stroke even after adjusting confounders in multivariate analysis. A hypertensive person was 3.161 times more likely to have ischemic stroke than a non-hypertensive person.

Some studies have shown that increased CCA-IMT correlates with the occurrence of ischemic stroke.⁵⁰ Lee et al. (2007) indicated that increased CCA-IMT was associated with the occurrence of ischemic stroke.⁵¹ Kumar et al. also pointed out that CCA-IMT was associated with the incidence of stroke, as well as its subtypes, and may be a predictor for the incidence of stroke,⁵² while some studies indicated that CCA-IMT was not associated with the risk of ischemic stroke.⁵³ Now the few existing studies did not explain the direct cause of ischemic stroke that caused by CCA-IMT thickening, while according to previous studies, it had been proved that CCA-IMT was related to the development of atherosclerosis, and CCA-IMT as the marker of atherosclerosis was associated with the incidence of ischemic stroke, carotid atherosclerosis had a positive association with the incidence of ischemic stroke.⁵⁴ Hyperlipidemia, hypertension, diabetes and so on can cause thickening of the CCA-IMT and lead to inflammation and endothelial dysfunction. In hyperlipidemia and other chronic diseases such as hypertension and diabetes, inflammatory factors and endothelium-related factors were increased, metabolism was abnormal, which result in inflammatory response and endothelial dysfunction, leading to CCA-IMT thickening and the development and aggravation of atherosclerotic lesions. Inflammation and endothelial dysfunction are related to the occurrence of atherosclerosis and ischemic stroke. Inflammation induces the proliferation and migration of vascular smooth muscle cells, which participates in the regulation of vascular tone and thus affects the development of atherosclerosis. Immune cells accumulate in intimal neovascularization and inhibit urokinase activation, leading to endothelial damage.⁵⁵ CCA-IMT gradually thickening can lead to cerebral artery stenosis or occlusion, can cause cerebral ischemia, lead to ischemic stroke. The thickening of CCA-IMT also can lead to carotid plaque formation, carotid plaque is instable, plaque is easy to fall off or easy to form thrombus, plaque falling off or thrombus falling off can block cerebral blood vessels and cause ischemic stroke. Our results showed that right CCA-IMT was an independent risk factor for ischemia. The incidence of ischemic stroke increased with an increase in right CCA-IMT; for every unit increase in the thickness of the right common carotid artery intima media, the odds of ischemic stroke increased by 45.273. In our study, right CCA-IMT and left CCA-IMT both had a relationship with ischemic stroke in the univariate binary logistic regression, and right CCA-IMT was associated with the incidence of ischemic stroke after adjusting for confounders in the multivariate analysis. While a significant relationship between left CCA-IMT and ischemic stroke was not present after the stepwise method in the multivariate analysis, only the right CCA-IMT still remained significant. Right CCA-IMT had more significant association with the incidence of ischemic stroke compared to left CCA-IMT. Because a few data of CCA-IMT was filled up in this study, this may affect the results slightly. In addition, there were almost no studies that could be referred for our results and proved the difference of right CCA-IMT and left CCA-IMT in the association with the incidence of ischemic stroke. We could not determine the reason that right CCA-IMT was associated with the incidence of ischemic stroke but left CCA-IMT not. Whether it is because there was a real difference between them or because of our sample reason, this needs more research to verify.

Conclusion

This study inferred that increased levels of Hcy and FBS and increased right CCA-IMT were associated with an increased incidence of ischemic stroke. Hcy, FBS, right CCA-IMT, and hypertension are independent risk factors for ischemic stroke in hospitalized patients in China. There was no association between TBIL, Scr, P-LCR, MPV, RDW, and the incidence of ischemic stroke. There was a sex difference in the association between MPV and ischemic stroke; increased MPV levels were found to be associated with the incidence of ischemic stroke only in females. There was no sex difference in the association between Hcy, TBIL, Scr, P-LCR, and the incidence of ischemic stroke. Targeted interventions and preventive measures for groups with a high risk of ischemic stroke could alleviate the burden of this disease. Nationwide screening of risk factors associated with ischemic stroke in middle-aged and elderly people as well as appropriate intervention studies are strongly recommended.