Sage Journals: Discover world-class research

Abstract

The role of metabolic syndrome (MetS) on stroke has been explored only in many observational studies. We conducted Mendelian randomization (MR) to clarify whether or not the genetically predicted MetS and its components are causally associated with stroke and its subtypes. Genetic instruments of MetS and its components and outcome data sets for stroke and its subtypes came from the gene-wide association study in the UK Biobank and MEGASTROKE consortium, respectively. Inverse variance weighting was utilized as the main method. Genetically predicted MetS, waist circumference (WC), and hypertension increase the risk of stroke. WC and hypertension are related to increased risk of ischemic stroke. MetS, WC, hypertension, and triglycerides (TG) are causally associated with the increasing of large artery stroke. Hypertension increased the risk of cardioembolic stroke. Hypertension and TG lead to 77.43- and 1.19-fold increases, respectively, in small vessel stroke (SVS) risk. The protective role of high-density lipoprotein cholesterol on SVS is identified. Results of the reverse MR analyses show that stroke is related to hypertension risk. From the genetical variants perspective, our study provides novel evidence that early management of MetS and its components are effective strategies to decrease the risk of stroke and its subtypes.

Keywords

Bidirectional causal association metabolic syndrome Mendelian randomization stroke

Introduction

The Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2019 reported that stroke ranks as the second leading cause of death and third leading cause of death and disability combined worldwide.¹ In 2019, the number of cases was 12.2 million, and the related deaths were 6.55 million. However, the disease burden has remained heavy despite substantial breakthroughs in emergency treatment for stroke since 2015.² Therefore, stroke prevention is regarded as an effective measure, given that 85% of all strokes may be preventable.³ At present, apart from non-intervenable factors (e.g., age and race), modifiable risk factors in life (e.g., tobacco use, alcohol consumption, and total cholesterol) gradually attract increasing interest in stroke prevention, given that it has substantially decreased the incidence of stroke by about 42% in high-income countries within the last 30 years.⁴ Thus, the high incidence of stroke may be decreased by identifying and through intervening modifiable risk factors.

The World Health Organization (WHO) defines metabolic syndrome (MetS) as a cluster of pathologic conditions, consisting of glucose abnormalities, hyperlipidemia, central obesity, and hypertension.⁵ In general, incidence of MetS is increasing rapidly, and approximately one quarter of people in Western countries have MetS.^6,7 Mounting evidence from many observational studies has indicated that MetS is independently correlated with the increased risk of stroke.⁸ However, clear confounders, such as retrospective studies and research designs, are inherent shortcomings in these studies, which may impede the understanding of the corresponding conclusions.

Mendelian randomization (MR) is a robust statistical approach to genetic variants to make a causal inference, which could overcome the limitation of observational studies.⁹ Single-nucleotide polymorphisms (SNPs) are assorted randomly in forming a zygote during gestation.¹⁰ Collectively, random assortment of interventions in MR analyses is similar to assignments in randomized clinical trials.^10,11 Moreover, no study has been conducted to investigate the causal association of MetS and its five components on stroke subtypes. Therefore, we performed this bidirectional MR analysis to yield their causal links.

Methods

Study design

The outline of our bidirectional MR study is shown in Figure 1. In the MR analysis, we identified the causal relationship between MetS and its five components and stroke and its subtypes. In addition, the causal effect of genetically predicted stroke and its subtypes on MetS and its components is investigated through reverse MR analysis. Ethical review and approval were not required in our study. Informed consent was acquired from all subjects in the original genome-wide association studies.

Figure 1.

The flow chart of our MR analysis. MR: mendelian randomization; SNP: single nucleotide ploymorphism; HDL-C: high-density lipoprotein cholesterol.

Date sources of exposures and outcomes

Genetic predictors for MetS and its components (waist circumference [WC], hypertension, triglycerides [TG], and high-density lipoprotein cholesterol [HDL-C]) came from the meta-analysis of gene-wide association study (GWAS) in the UK Biobank (MetS: 291,107 participants; WC: 462,166 samples; hypertension: 463,010 subjects; TG: 441,016 samples; HDL-C: 403,943 participants).^12–14 Instrumental variables (IVs) of fasting blood glucose (FBG) were available from GWAS in the meta-analyses of glucose and insulin-related traits consortium (MAGIC), consisting of 281,416 samples.¹⁵ The detailed sources of data sets used in our study are summarized in Supplementary Table 1.

Outcome data sets on stroke and its subtypes were taken from the MEGASTROKE consortium in the European population.¹⁶ The data sets consisted of five stroke traits, (stroke: 67,162 cases and 453,702 control; ischemic stroke [IS]: 60,341 cases and 452,221 controls; large artery stroke [LAS]: 6688 cases and 238,763 controls; cardioembolic stroke [CES]: 9006 cases and 378,720 controls; and small vessel stroke [SVS]: 11,710 cases and 286,854 controls). Stroke is diagnosed based on WHO’s definition, which is rapidly exacerbating signs of focal (or global) cerebral dysfunction, experiencing over 24 hours, or resulting in death without evident triggers other than that of the blood vessel source.

Genetic instrument selection

Genetic instruments related to MetS and its components should meet the criterion that all SNPs satisfied a significance level (P < 5 × 10⁻⁸), and an r² threshold was <0.05 when we pruned independent SNPs in linkage disequilibrium analysis (at a window size of 10,000 kb). The MR Pleiotropy RESidual Sum and Outlier (MR-PRESSO) approach was employed to identify significant outliers explaining possible pleiotropy,¹⁷ which should be removed to reduce the heterogeneity and horizontal pleiotropy. Value of the F-statistics represents the strength of our MR study, and the computational formula is F-statistics = (Beta/Se),² where Beta and Se are the relational coefficients between IVs and traits and standard error, respectively. Value of F-statistics >10 is an indicator of statistical robustness.¹⁸ Lastly, qualified SNPs selected as IVs are presented in Table 1.

Table 1.

The R² and F-statistics for the genetic instruments in the MR and reverse MR analyses.

Exposure	Outcome	No. SNP	R²	F-statistic
Mets	Stroke	127	3.27%	68.31
WC	Stroke	567	7.27%	53.88
Hypertension	Stroke	65	0.84%	45.57
FBG	Stroke	107	4.35%	101.59
TG	Stroke	751	19.31%	121.77
HDL-C	Stroke	903	29.80%	162.78
Mets	IS	127	3.27%	68.31
WC	IS	566	7.29%	54.01
Hypertension	IS	61	0.78%	45.32
FBG	IS	107	4.35%	101.59
TG	IS	754	19.37%	121.47
HDL-C	IS	902	29.86%	163.09
Mets	LAS	124	3.19%	68.13
WC	LAS	569	7.32%	54.06
Hypertension	LAS	62	0.81%	45.89
FBG	LAS	107	4.35%	101.59
TG	LAS	759	19.04%	118.58
HDL-C	LAS	904	29.78%	162.32
Mets	CES	125	3.23%	68.93
WC	CES	572	7.34%	53.99
Hypertension	CES	65	0.84%	46.12
FBG	CES	107	4.35%	101.59
TG	CES	757	19.19%	119.75
HDL-C	CES	904	29.76%	162.27
Mets	SVS	127	3.27%	68.31
WC	SVS	569	7.27%	53.83
Hypertension	SVS	62	0.78%	44.79
FBG	SVS	107	4.35%	101.59
TG	SVS	755	19.39%	121.60
HDL-C	SVS	903	29.63%	161.45
Stroke	Mets	5	0.42%	41.01
Stroke	WC	6	0.49	41.98
Stroke	Hypertension	4	0.36%	41.54
Stroke	FBG	5	0.44%	42.57
Stroke	TG	4	0.37%	41.83
Stroke	HDL-C	3	0.30%	41.34
IS	Mets	7	0.51%	38.56
IS	WC	4	0.32%	39.08
IS	Hypertension	3	0.27%	42.32
IS	FBG	7	0.54%	40.99
IS	TG	5	0.40%	40.51
IS	HDL-C	4	0.32%	39.63
LAS	Mets	–	–	–
LAS	WC	3	2.21%	37.89
LAS	Hypertension	–	–	–
LAS	FBG	3	2.21%	37.89
LAS	TG	3	2.21%	37.89
LAS	HDL-C	3	2.21%	37.89
CES	Mets	4	3.25%	75.64
CES	WC	5	3.67%	68.58
CES	Hypertension	5	3.67%	68.58
CES	FBG	5	3.67%	68.58
CES	TG	5	3.67%	68.58
CES	HDL-C	3	2.68%	82.75
SVS	Mets	–	–	–
SVS	WC	–	–	–
SVS	Hypertension	–	–	–
SVS	FBG	–	–	–
SVS	TG	–	–	–
SVS	HDL-C	–	–	–

IS: ischemic stroke; LAS: large artery stroke; CES: cardioembolic stroke; SVS: small vessel stroke; MetS: metabolic syndrome; WC: waist circumference; FBG: fasting blood glucose; TG: triglycerides; HDL-C: high-density lipoprotein cholesterol.

Main statistical analyses

The main analytical method is the random effects inverse variance weighting (re-IVW) approach to identify the causal effect of risk factors on outcomes in the MR and reverse MR analyses. The reason is that the IVW method can obtain a robust result in the condition of no signs of pleiotropy (no violation of the independence assumption).

Statistical significance is set at the threshold of the Bonferroni-corrected P < 0.0016 (0.05/30) in this study. All analyses were conducted using the “TwoSampleMR”, “mr.raps”, “cause”, and “forestplot” packages in R software (version: 3.6.5).

Sensitivity analyses

This study used five methods to perform sensitivity analyses. The comparison result of egger intercept term with zero in the MR-Egger analysis represents the directional pleiotropy. Even if all IVs are invalid, MR-Egger can obtain valid causal effect estimates.¹⁹ The weighted median approach can gain consistent findings via aggregating numerous estimates into a single total estimate, even if half of SNPs are not valid.²⁰ A relative low standard error exists in the maximum likelihood method, although it may be deviated by a small sample.²¹ The findings in the MR-RAPS analysis remain robust in the presence of weak IVs, on the condition that horizontal pleiotropy exists.²² The MR-PRESSO method can identify significant outliers and remove them thereafter for pleiotropy. Collectively, markedly reliable results can be produced.¹⁷ The results of the global test were used to account for horizontal pleiotropy.

The egger intercept term in the MR-Egger analysis and P value of the MR-PRESSO approach were introduced into the regression model to detect directional pleiotropy. We also performed Cochran’s Q test to determine possible heterogeneity. In addition, the findings of the leave-one-out analysis was used to illustrate the robustness of the conclusion when excluding IVs at a time.

Results

Causal effect of genetically predicted MetS and its components on stroke

Results from the IVW analysis are presented in Table 2, and suggest that MetS increases a 1.05-fold risk of stroke (95% confidence interval [CI]: 1.02–1.09, P = 0.0012). Scatter plot (Figure 2) shows that the IV effect on stroke is intensified when the SNP effect on MetS increases. In sensitivity analyses, causal association between MetS and stroke still exists (Maximum likelihood method: OR: 1.06, 95%CI: 1.02–1.09, P = 0.0002; MR-RAPS: OR: 1.06, 95%CI: 1.02–1.10, P = 0.0011; Table 2). Results in the Cochran’s Q analysis demonstrate visible heterogeneity (Q = 163.13, P = 0.01), while the funnel plot in Figure 3 shows a symmetry of the MR results, suggesting no signs of heterogeneity. No pleiotropy was identified using the MR-Egger (egger intercept term = 0.0040, P = 0.112) and MR-PRESSO methods (P > 0.05). Moreover, no influential IVs were identified in the leave-one-out analysis when excluding any one of SNPs (Figure 4). Frost plot manifesting the causal estimate of every SNP on stroke is shown in Figure 5.

Table 2.

The causal effect of MetS and its components on stroke and its subtypes.

Exposure	Outcome	Methods	OR (95%CI)	P	Egger_intercept	P-Egger_intercept
MetS	Stroke	IVW	1.05 (1.02,1.09)	0.0012
		MR-Egger	0.99 (0.92,1.08)	0.9754	0.0040	0.112
		Weighted median	1.05 (1.00,1.10)	0.0316
		Maximum likelihood	1.06 (1.02,1.09)	0.0002
		RAPS	1.06 (1.02,1.10)	0.0011
WC	Stroke	IVW	1.23 (1.15,1.32)	2.39e-09
		MR-Egger	1.22 (0.99,1.50)	5.80e-02	0.0001	0.915
		Weighted median	1.19 (1.06,1.34)	1.82e-03
		Maximum likelihood	1.23 (1.16,1.31)	1.26e-11
		RAPS	1.23 (1.14,1.32)	1.27e-08
Hypertension	Stroke	IVW	47.99 (25.01,92.10)	2.54e-31
		MR-Egger	15.51 (1.48,162.69)	2.55e-02	0.0060	0.330
		Weighted median	43.07 (19.21,96.55)	6.36e-20
		Maximum likelihood	54.05 (31.56,92.57)	7.07e-48
		RAPS	51.91 (27.04,99.64)	1.65e-32
FBG	Stroke	IVW	1.09 (0.95,1.23)	0.1861
		MR-Egger	0.89 (0.70,1.12)	0.3542	0.0051	0.051
		Weighted median	0.92 (0.77,1.10)	0.3932
		Maximum likelihood	1.09 (0.98,1.21)	0.0991
		RAPS	1.06 (0.92,1.21)	0.3802
TG	Stroke	IVW	1.06 (1.02,1.10)	0.0012
		MR-Egger	0.99 (0.93,1.05)	0.7947	0.0022	0.004
		Weighted median	1.05 (0.98,1.12)	0.1428
		Maximum likelihood	1.06 (1.02,1.10)	0.0003
		RAPS	1.06 (1.02,1.11)	0.0015
HDL_C	Stroke	IVW	0.94 (0.91,0.98)	0.0017
		MR-Egger	1.01 (0.96,1.06)	0.6121	0.0007	0.001
		Weighted median	0.96 (0.91,1.02)	0.2670
		Maximum likelihood	0.94 (0.91,0.97)	0.0007
		RAPS	0.94 (0.91,0.98)	0.0022

MetS: metabolic syndrome; WC: waist circumference; FBG: fasting blood glucose; TG: triglycerides; HDL-C: high-density lipoprotein cholesterol; IVW: inverse-variance weighted; RAPS: robust adjusted profile score; OR: odds ratio; CI: confidential interval.

Figure 2.

The scatter plots of the association between genetically predicted MetS and its components on stroke in the MR analysis. IS: ischemic stroke; LAS: large artery stroke; CES: cardioembolic stroke; SVS: small vessel stroke; HDL-C: high-density lipoprotein cholesterol; IVW: inverse-variance weighted; RAPS: robust adjusted profile score.

Figure 3.

The funnel plots of the association between genetically predicted stroke on MetS and its components in the reverse MR analysis. IS: ischemic stroke; LAS: large artery stroke; CES: cardioembolic stroke; SVS: small vessel stroke; HDL-C: high-density lipoprotein cholesterol; IVW: inverse-variance weighted.

Figure 4.

The leave-one-out analysis of the association between genetically MetS and its components on stroke in the MR analysis. IS: ischemic stroke; LAS: large artery stroke; CES: cardioembolic stroke; SVS: small vessel stroke; HDL-C: high-density lipoprotein cholesterol; OR: odds ratio.

Figure 5.

The frost plots of the association between genetically MetS and its components on stroke in the MR analysis. IS: ischemic stroke; LAS: large artery stroke; CES: cardioembolic stroke; SVS: small vessel stroke; HDL-C: high-density lipoprotein cholesterol; OR: odds ratio.

For MetS’ components, WC and hypertension were genetically related to the increasing risk of stroke, while not related to FBG, TG, and HDL-C (all P > 0.0016, Table 2, Figure 2). ORs with 95%CI on stroke is 1.23 (95%CI: 1.15–1.32, P = 2.39 × 10⁻⁹) for WC and 47.99 (95%CI: 25.01–92.10, P = 2.54 × 10⁻³¹) for hypertension. The causal association of HDL-C on stroke is suggestive (P = 0.0017), while the protective effect of HDL-C on stroke is identified (OR: 0.94, 95%CI: 0.91–0.98). Nevertheless, potential horizontal pleiotropy is observed for TG and HDL-C in the MR-Egger regression analyses (P < 0.05), and the causal effect of TG and HDL-C on stroke estimated by CAUSE is unsupported (P = 0.56, P = 0.33, respectively). Evident heterogeneity is detected in Cochran’s Q analysis (all P < 0.05), but MR results in the funnel plot (Figure 3) are symmetrical, suggesting the absence of heterogeneity. In addition, the results in the leave-one-out analysis reveal a robust positive association after leaving any single SNP out in turn (Figure 4). This result indicates that no influential SNPs are found. Causal estimates for every IV are shown in Figure 5.

Causal effect of genetically predicted MetS and its components on is-IS

Harmful effect of genetically predicted MetS on IS has weak statistical significance (OR: 1.05, 95%CI: 1.01–1.09, P = 0.0032), which is similar to the results in the MR-RAPS analysis (OR: 1.06, 95%CI: 1.02–1.10, P = 0.0021) (Supplementary Table 2). As shown in Figure 2, the risk of IS tends to increase as IV’s effect on MetS increases. Although results of Cochran’s Q analysis suggest heterogeneity, MR results in the funnel plot are symmetrical (Figure 3). This result indicates an absence of heterogeneity. Accordingly, there are no evidence of pleiotropy in MR-Egger (egger intercept term = 0.0051, P = 0.082) and MR-PRESSO analyses (P > 0.05). Moreover, no influential IVs are identified in the leave-one-out test after omitting any SNPs (Figure 4). The results of the MR analyses for each SNP are presented in Figure 5.

For the components (Figure 2), the risk of IS is increased by WC (OR: 1.25, 95%CI: 1.16–1.34, P = 2.03 × 10⁻⁹) and hypertension (OR: 59.54, 95%CI: 32.64–108.60, P = 1.57 × 10⁻⁴⁰). TG may increase the risk of IS, but causal association is not supported and estimated by CAUSE (P = 0.94). FBG and HDL-C are not related to increased risk of IS after performing CAUSE analyses (all P > 0.05). For heterogeneity, Cochran’s Q analysis indicates evident heterogeneities for WC, FBG, TG, and HDL-C (all P < 0.05). Results in all funnel plots suggest no signs of heterogeneity (Figure 3). Meanwhile, there is absence of pleiotropy in the MR-Egger (all P > 0.05) and MR-PRESSO (all P > 0.05) analyses. Furthermore, results of the leave-one-out analyses suggest that no influential IVs are identified after removing any SNP in turn (Figure 4). Estimates for every SNP are shown in Figure 5.

Causal effect of genetically predicted MetS and its components on LAS

As shown in the Supplementary Table 3, genetically predicted MetS is positively related to LAS (OR: 1.21, 95%CI: 1.10–1.33, P = 2.96 × 10⁻⁵). Positive causal association remains supported in the maximum likelihood (OR: 1.21, 95%CI: 1.11–1.32, P = 5.79 ×10⁻⁶) and MR-RAPS (OR: 1.20, 95%CI: 1.09–1.32, P = 1.50 × 10⁻⁴) analyses. Figure 2 shows the harmful effect of MetS on LAS. In Cochran’s Q test, no sign of heterogeneity is detected (P values of Cochran’s Q > 0.05, as shown in Figure 3). No signs of pleiotropy are observed in the MR-Egger (egger intercept term = 0.0101, P = 0.157) and MR-PRESSO (P > 0.05) analyses. In addition, results in the leave-one-out analysis demonstrate that the causal association is robust after leaving any single SNP out in turn (Figure 4). The forest plot manifesting the estimate of every SNP on LAS is shown in Figure 5.

Positive causal associations among WC, hypertension, TG, and LAS are also observed. ORs with 95%CI on LAS is 1.47 (95%CI: 1.23–1.76, P = 1.91 × 10⁻⁵) for WC, 25.05 (95%CI: 5.07–123.80, P = 7.75 × 10⁻⁵) for hypertension, and 1.22 (95%CI: 1.10–1.36, P = 1.46 × 10⁻⁴) for TG (Figure 2). Causal effect of genetically predicted FBG on LAS is on the edge of significance (P = 0.0017), and FBG has a trend to increase the risk of LAS (OR: 1.62, 95%CI: 1.20–2.20). Despite pleiotropy, there are no signs of causal association between HDL-C and LAS after being estimated by CAUSE (P = 0.079). Heterogeneity of WC, hypertension, TG, and HDL-C is detected in Cochran’s Q test, although the symmetry of MR results in the funnel plots suggests no sign of heterogeneity (Figure 3). Causal association among WC, hypertension, TG, and LAS remains stable after removing any SNP in turn (Figure 4). Figure 5 presents the estimates of each SNP on LAS.

Causal effect of genetically predicted MetS and its components on CES

Results of IVW demonstrate that MetS and its five components (WC, hypertension, FBG, TG, and HDL-C) are not significantly correlated to CES (all P > 0.05), while hypertension increases the risk of CES (OR: 55.51, 95%CI: 15.52–198.53, P = 6.48 ×10⁻¹⁰) (Supplementary Table 4). Causal association of MetS and its five components on CES visualizing in the scatter plots reveals that MetS, WC, FBG, TG, and HDL-C do not increase the risk of CES (Figure 2). Moreover, there is no sign of heterogeneity for MetS, WC, FBG, and TG according to findings of Cochran’s Q analyses (P > 0.05). Although heterogeneity for hypertension and HDL-C is detected in the Cochran’s Q analyses (P < 0.05), the funnel plots (Figure 3) suggest a symmetry of the MR results, indicating no signs of heterogeneity. No evidence of pleiotropy is observed in the MR-Egger analyses for MetS and its components (all P > 0.05) and MR-PRESSO analyses (all P > 0.05). The leave-one-out analysis obtains consistent results when leaving any SNP out in turn (Figure 4). Frost plots display the causal estimates from each IV on CES (Figure 5).

Causal effect of genetically predicted MetS and its components on SVS

As shown in Supplementary Table 5, a potential causal association of MetS and WC on SVS is observed (MetS: OR: 1.09, 95%CI: 1.00–1.19; P = 0.0441; WC: OR: 1.18, 95%CI: 1.01–1.39; P = 0.0471). Moreover, causal effect of genetically predicted hypertension and TG on SVS is significant (hypertension: OR: 77.43, 95%CI: 19.48–307.64; P = 6.43 × 10⁻¹⁰; TG: OR: 1.19, 95%CI: 1.08–1.31, P = 3.44 × 10⁻⁴). Causal association between FBG and SVS does not exist after performing CAUSE analysis (P > 0.05). HDL-C may be a protective factor for SVS (OR: 0.84, 95%CI: 0.77–0.92, P = 1.41 × 10⁻⁴). Scatter plots reveal that hypertension and TG increase the risk of SVS (Figure 2). Results of MetS, WC, FBG, TG, and HDL-C in Cochran’s Q tests reveal visible heterogeneity (all P values of Cochran’s Q test < 0.05). However, funnel plots (Figure 3) present a symmetrical MR analysis, indicating an absence of heterogeneity. No evidence of pleiotropy is observed in the MR-Egger analyses for MetS, WC, hypertension, TG, and HDL-C (all P > 0.05) and MR-PRESSO analyses (P > 0.05). Causal association between FBG and SVS is unsupported after being estimated by CAUSE (P = 0.65). In addition, the leave-one-out analysis suggests a robust causal association after removing any SNP in turn (Figure 4). Causal estimates for every IV on SVS are shown in Figure 5.

Causal effect of genetically predicted stroke on MetS and its components in the reverse MR analysis

Results of the IVW analysis (Supplementary Table 6–9 and Supplementary Figure 1) show that neither stroke nor IS, LAS, CES, and SVS are causally related to MetS and its five elements (all P > 0.05), except the association between stroke and hypertension. Stroke may have harmful effects on hypertension (OR: 1.03, 95%CI: 1.01–1.06, P = 2.13 × 10⁻⁴). Results of the MR-Egger and MR-PRESSO analyses reveal that no potential pleiotropy exists (all P > 0.05). Cochran’s Q analyses (all P > 0.05) and MR results in the funnel plots (Supplementary Figure 2) do not detect heterogeneity. Results of the leave-one-out analyses also remain stable (Supplementary Figure 3). Causal estimates from each IV on SVS are shown in Supplementary Figure 4.

Discussion

In our bidirectional MR analyses, we strengthen the evidence that MetS and its five components are important risk factors for stroke and its subtypes. Results in the reverse MR analyses reveal that liability to stroke is not associated with MetS and its components, even though stroke increases the risk of hypertension.

Although the definition or diagnostic of MetS have not been harmonized, it is widely recognized that MetS increases the risk of stroke. Epidemiological evidence has proposed that MetS is significantly correlated with increased risk of stroke^8,23 For example, results of a 9.1-year follow-up of 5,171 subjects indicate that people with MetS had higher risk of stroke compared with those without (hazard ratio (HR) = 1.86).²³ Findings of a meta-analysis, including 16 cohorts with 116,496 samples, reveal that MetS is regarded as an independent hazard factor for stroke.²⁴ In our MR analyses, results supported the causal association between MetS and stroke. Moreover, research on the causal association of MetS on LAS, CES, and SVS is scant at present. This MR study first provides evidence that MetS increases the risk of LAS but not CES. MetS is significantly related to increased risk of SVS. For the reverse association, our MR results show that causal association between stroke and MetS does not exist.

For the elements of MetS, two large cohorts reported that WC is tightly correlated with increased risk of stroke among Westerners.^25,26 However, evidence on association was inconsistent. A study in Japan only found positive relationships of WC within women.²⁷ Inconsistent results may be a result of race, lifestyle, severity of obesity, or some significant risk factors. In our MR analysis, we found that WC is closely associated with the risk of stroke, IS, and LAS, while a potential causal association is detected between WC and SVS and CES. Accumulating evidence suggests that higher FBG level significantly increases the risk of stroke and may be regarded as a predictor for stroke.^28–31 A comprehensive meta-analysis of 102 observational studies suggests that FBG is closely associated with increased risk of stroke.²⁹ However, no significant relationship between higher FBG levels and stroke was identified in our MR analysis. Moreover, causal association between EBG and LAS is suggestive. Circulating lipoprotein was significantly associated with stroke.³² Observational studies have suggested inconsistent results on the association among TG, HDL-C, and stroke; most of them have supported a direct relationship.^33,34 By contrast, research on subtypes of stroke has been scant in most observational studies. An MR study has suggested that HDL-C is a protective factor for SVS, which is supported by our MR results.³⁵ Moreover, the current study does not support the causal association of genetically high TG on stroke and its subtypes.³⁵ However, we found that TG is positively related to the risks of LAS and SVS. This inconsistent result may be a result of sample size and parameters. Therefore, future studies with large sample size are required to explore the role of TG on stroke.

Although the clear mechanisms linking MetS and its components to stroke remain unclear, several possible explanations have been proposed. Some MetS-related changes lead to damages in the fibrinolysis competence, hyperglycemia, endothelium injury, and pro-inflammatory response, thereby possibly leading to aggravating cerebrovascular injury and further resulting in stroke.³⁶ Moreover, inflammatory processes of atherosclerotic change in vessels may play a significant role in the risk for stroke in persons with distinct elements of MetS, such as FBG and WC. Inflammation response may lead to vascular degradation and insulin resistance³⁷ and the release of endothelin-1 caused by resisting results in the constriction of the artery after stroke.³⁸ Therefore, inflammatory processes strike a significant effect on the link of MetS on increased risk of stroke.

Results of previous observational studies may have been influenced by numerous potential confounders, such as study design, limited sample size, or retrospective study in clinics. The strength of the MR study overcomes the underlying impact of confounding factors and yields causal inferences. Moreover, we performed bidirectional MR analyses to facilitate the understanding of the results. In addition, we solved the pleiotropy in our MR analysis. However, this MR study still has some limitations. First, the participants of LAS, CES, and SVS are relatively small. Second, bias derived from population stratification could not be ruled out. Lastly, the data sets originated from the European population, thereby limiting the generalization of the conclusion. In future studies, replication in other ancestries, more rigorous clinical study design, and large studies with more samples should be performed to verify the conclusions.

Conclusion

Our study supports the causal association between MetS and its five components and stroke and its subtypes. In the reverse MR analysis, no association was detected except the link between stroke and hypertension. Early management of MetS and its elements should be regarded as important measures in the first-class prevention of stroke.

Supplemental Material

sj-pdf-1-jcb-10.1177_0271678X231169838 - Supplemental material for Genetic insights into the risk of metabolic syndrome and its components on stroke and its subtypes: Bidirectional Mendelian randomization

Supplemental material, sj-pdf-1-jcb-10.1177_0271678X231169838 for Genetic insights into the risk of metabolic syndrome and its components on stroke and its subtypes: Bidirectional Mendelian randomization by Qiang He, Wenjing Wang, Hao Li, Yang Xiong, Chuanyuan Tao, Lu Ma and Chao You in Journal of Cerebral Blood Flow & Metabolism

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the 1·3·5 project for disciplines of excellence-Clinical Research Incubation Project, West China Hospital, Sichuan University (2018HXFH010).

Data availability statement

All data in our MR analyses is available from public databases.

Acknowledgements

We give the great appreciation to the participants and working staff for their excellent job to the study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions

Q.H and W.W analyzed the data and wrote the manuscript; H.L, Y.X, and Q.H interpreted the data; C.T, L.M and C.Y designed the study.

Supplemental material

Supplemental material for this article is available online.

References

Collaborators GBDS. Global, regional, and national burden of stroke and its risk factors, 1990–2019: a systematic analysis for the global burden of disease study 2019. Lancet Neurol 2021; 20: 795–820.

Campbell

BCV

Khatri

Stroke. Lancet 2020; 396: 129–142.

O'Donnell

Xavier

Liu

, et al. Risk factors for ischaemic and intracerebral haemorrhagic stroke in 22 countries (the interstroke study): a case-control study. Lancet 2010; 376: 112–123.

Spence

JD.

Nutrition and risk of stroke. Nutrients 2019; 11: 647.

Eckel

Grundy

Zimmet

PZ.

The metabolic syndrome. Lancet 2005; 365: 1415–1428.

Mogre

Salifu

Abedandi

Prevalence, components and associated demographic and lifestyle factors of the metabolic syndrome in type 2 diabetes mellitus. J Diabetes Metab Disord 2014; 13: 80.

Xiong

Zhang

, et al. Prevalence and associated factors of metabolic syndrome in chinese Middle-aged and elderly population: a national cross-sectional study. Aging Male 2021; 24: 148–159.

Sarrafzadegan

Gharipour

Sadeghi

, et al. Metabolic syndrome and the risk of ischemic stroke. J Stroke Cerebrovasc Dis 2017; 26: 286–294.

Emdin

Khera

Kathiresan

Mendelian randomization. JAMA 2017; 318: 1925–1926.

10.

Smith

Ebrahim

‘ Mendelian randomization': can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 2003; 32: 1–22.

11.

Burgess

Butterworth

Thompson

SG.

Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 2013; 37: 658–665.

12.

Lind

Genome-wide association study of the metabolic syndrome in UK biobank. Metab Syndr Relat Disord 2019; 17: 505–511.

13.

Elsworth

Mitchell

Raistrick

, et al. UK: University of Bristol. Mrc Ieu uk Biobank Gwas Pipeline Version 1 2017.

14.

Richardson

Sanderson

Palmer

, et al. Evaluating the relationship between circulating lipoprotein lipids and apolipoproteins with risk of coronary heart disease: a multivariable mendelian randomisation analysis. PLoS Med 2020; 17: e1003062.

15.

Chen

Spracklen

Marenne

, et al. The trans-ancestral genomic architecture of glycemic traits. Nat Genet 2021; 53: 840–860.

16.

Malik

Chauhan

Traylor

, et al. Multiancestry genome-wide association study of 520,000 subjects identifies 32 loci associated with stroke and stroke subtypes. Nat Genet 2018; 50: 524–537.

17.

Verbanck

Chen

Neale

, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from mendelian randomization between complex traits and diseases. Nat Genet 2018; 50: 693–698.

18.

Chen

Yang

, et al. Insights into modifiable risk factors of cholelithiasis: a Mendelian randomization study. Hepatology 2022; 75: 785–796.

19.

Bowden

Davey Smith

Burgess

Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol 2015; 44: 512–525.

20.

Bowden

Davey Smith

Haycock

, et al. Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator. Genet Epidemiol 2016; 40: 304–314.

21.

Milligan

BG.

Maximum-likelihood estimation of relatedness. Genetics 2003; 163: 1153–1167.

22.

Zhao

Chen

Wang

, et al. Powerful three-sample genome-wide design and robust statistical inference in summary-data mendelian randomization. Int J Epidemiol 2019; 48: 1478–1492.

23.

Sanchez-Inigo

Navarro-Gonzalez

Fernandez-Montero

, et al. Risk of incident ischemic stroke according to the metabolic health and obesity states in the vascular-metabolic CUN cohort. Int J Stroke 2017; 12: 187–191.

24.

Fang

, et al. Is metabolic syndrome associated with the risk of recurrent stroke: a meta-analysis of cohort studies. J Stroke Cerebrovasc Dis 2017; 26: 2700–2705.

25.

Pischon

Boeing

Hoffmann

, et al. General and abdominal adiposity and risk of death in Europe. N Engl J Med 2008; 359: 2105–2120.

26.

Gelber

Gaziano

Orav

, et al. Measures of obesity and cardiovascular risk among men and women. J Am Coll Cardiol 2008; 52: 605–615.

27.

Furukawa

Kokubo

Okamura

, et al. The relationship between waist circumference and the risk of stroke and myocardial infarction in a Japanese urban cohort: the Suita study. Stroke 2010; 41: 550–553.

28.

Liu

Yang

, et al. Association of fasting glucose levels with incident atherosclerotic cardiovascular disease: an 8-year follow-up study in a Chinese population. J Diabetes 2017; 9: 14–23.

29.

Sarwar

Gao

Seshasai

SRK

, et al. Diabetes mellitus, fasting blood glucose concentration, and risk of vascular disease: a collaborative meta-analysis of 102 prospective studies. Lancet 2010; 375: 2215–2222.

30.

Liu

Rutten-Jacobs

Liu

, et al. Causal impact of type 2 diabetes mellitus on cerebral small vessel disease: a mendelian randomization analysis. Stroke 2018; 49: 1325–1331.

31.

Luitse

Biessels

Rutten

, et al. Diabetes, hyperglycaemia, and acute ischaemic stroke. Lancet Neurol 2012; 11: 261–271.

32.

Yaghi

Elkind

MS.

Lipids and cerebrovascular disease: research and practice. Stroke 2015; 46: 3322–3328.

33.

Shahar

Chambless

Rosamond

, et al. Plasma lipid profile and incident ischemic stroke: the atherosclerosis risk in communities (ARIC) study. Stroke 2003; 34: 623–631.

34.

Labreuche

Touboul

Amarenco

Plasma triglyceride levels and risk of stroke and carotid atherosclerosis: a systematic review of the epidemiological studies. Atherosclerosis 2009; 203: 331–345.

35.

Hindy

Engstrom

Larsson

, et al. Role of blood lipids in the development of ischemic stroke and its subtypes: a mendelian randomization study. Stroke 2018; 49: 820–827.

36.

Arenillas

Moro

Davalos

The metabolic syndrome and stroke: potential treatment approaches. Stroke 2007; 38: 2196–2203.

37.

Berg

Scherer

PE.

Adipose tissue, inflammation, and cardiovascular disease. Circ Res 2005; 96: 939–949.

38.

Wang

Ruotsalainen

Moilanen

, et al. The metabolic syndrome predicts incident stroke: a 14-year follow-up study in elderly people in Finland. Stroke 2008; 39: 1078–1083.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.11 MB