If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence to: L. Wang, Department of Epidemiology and Biostatistics, Institute of Basic Medical Sciences Chinese Academy of Medical Sciences; School of Basic Medicine Peking Union Medical College, 5 Dong Dan San Tiao, Beijing, 100005, China.
Department of Epidemiology and Biostatistics, Institute of Basic Medical Sciences Chinese Academy of Medical Sciences; School of Basic Medicine Peking Union Medical College, 5 Dong Dan San Tiao, Beijing, 100005, China
Mendelian randomization was performed for the causal effect of NAFLD on CVD events.
•
The study strongly supported the causal effect of NAFLD on arterial stiffness.
•
In the study, NAFLD was not causally associated with CAD, heart failure, stroke, and its subtypes.
Abstract
Background
Evidence suggests that nonalcoholic fatty liver disease (NAFLD) is associated with cardiovascular diseases (CVDs). However, the results are inconsistent, and the causality remains to be established.
Objective
We aimed to investigate the potential causal relationship between NAFLD and CVDs, including arterial stiffness, coronary artery disease, heart failure, stroke, ischemic stroke and its subtypes using two-sample Mendelian randomization (MR).
Methods
Genetic instruments were used as proxies for NAFLD. Publicly available summary-level data were obtained from the UK Biobank, the CARDIoGRAMplusC4D Consortium, the MEGASTROKE Consortium, and other consortia. Six complementary MR methods were performed, including inverse variance weighted method (IVW), MR-Egger, weighted median, weighted mode, MR-PRESSO, and MR-RAPS.
Results
NAFLD was significantly associated with arterial stiffness (β = 0.04 [95%CI, 0.02–0.06], P = 5.53E-04). Moreover, the results remained consistent and robust in the sensitivity analysis. As for heart failure, the IVW method suggested that NAFLD was significantly associated with heart failure (OR = 1.08, 95%CI: 1.02–1.14, P = 0.005) in the absence of pleiotropy. However, there were no significant associations of NAFLD with coronary artery disease, stroke, ischemic stroke, or any ischemic stroke subtype.
Conclusion
The MR study supported the causal effect of NAFLD on arterial stiffness. However, the study did not provide enough evidence suggesting the causal associations of NAFLD with heart failure, coronary artery disease, and any stroke subtypes.
]. Recently, researchers have stimulated a growing interest in the potential role of NAFLD in developing cardiovascular diseases (CVDs) by increasing awareness of the significance of NAFLD and its growing prevalence. Accumulating evidence from epidemiological studies showed that NAFLD was associated with an increased risk of CVD events. A nationwide histology cohort reported that biopsy-proven NAFLD could significantly increase the risk of major adverse cardiovascular events and increase with worsening disease severity, including ischemic heart disease, stroke, congestive heart failure, or cardiovascular mortality [
]. A meta-analysis confirmed that NAFLD was associated with an increased risk of myocardial infarction (MI), ischemic stroke (IS), atrial fibrillation (AF), and heart failure (HF) [
However, most of the evidence from observational studies was not entirely consistent. For example, a large-scale cohort study of 18 million European adults found no significant association of NAFLD with acute myocardial infarction or stroke risk [
Non-alcoholic fatty liver disease and risk of incident acute myocardial infarction and stroke: findings from matched cohort study of 18 million European adults.
]. The conventional observational studies were inevitably subject to various confounding factors and reverse causation bias. Therefore, whether NAFLD would causally contribute to the etiology of CVD is still unknown.
Furthermore, though several studies examined the association between NAFLD and stroke, they had mainly focused on ischemic or unspecified stroke [
Non-alcoholic fatty liver disease and risk of incident acute myocardial infarction and stroke: findings from matched cohort study of 18 million European adults.
] while sufficiently strong evidence in different stroke subtypes is still lacking. The heterogeneity of stroke subtypes might potentially influence the association between NAFLD and stroke. Therefore, it is imperative to investigate the association between NAFLD and different ischemic stroke subtypes, including large artery stroke (LAS), cardioembolic stroke (CES), and small vessel stroke (SVS).
The rapid development of genome-wide association studies (GWAS) has led to the increasing application of Mendelian randomization (MR) analysis using the phenotypic-associated SNPs as instrumental variables (IVs) [
], it is unethical and inappropriate to perform RCT on the current topic. On the contrary, the MR method allows examining causal association regardless of research costs and ethics. Moreover, the alleles used as IVs were randomly allocated at conception, making the MR method a “natural randomised control trial” [
]. Therefore, MR is sometimes an excellent and powerful tool for causal inference.
In summary, we performed the two-sample MR analyses to investigate the causal association of NAFLD with different CVD events, including arterial stiffness, coronary artery disease, heart failure, stroke, ischemic stroke, and its subtypes (LAS, CES, SVS).
2. Methods
2.1 Two-sample MR
Two-sample MR was performed in the analysis. Single nucleotide polymorphisms (SNPs) were randomly allocated at conception, determining that SNPs were not affected by post environmental factors [
]. The contribution of widely growing open databases facilitates the application of the two-sample MR studies. All the summary data used in the study are publicly available, and the detailed information is presented in Table 1.
Table 1Details of data source included in the study.
M Vujkovic S Ramdas KM Lorenz , et al. A trans-ancestry genome-wide association study of unexplained chronic alt elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. medRxiv. 20212020.12.26.20248491. doi:10.1101/2020.12.26.20248 491.
(i) Elevated ALT >40 U/L for men and >30 U/L for women during at least two time points at least 6 months apart within a two-year window period at any point prior to enrollment; (ii) exclusion of other causes of liver disease.
Age, gender, age-adjusted AUDIT-C score, and 10 principal components of genetic ancestry.
HF cases were included with a clinical diagnosis of HF of any etiology with no inclusion criteria based on LV ejection fraction; controls were participants without HF.
The presence of myocardial infarction, acute coronary syndrome, chronic stable angina, or coronary stenosis >50%, and this definition could vary among the 48 studies included in the meta-analysis.
Sex, age, and generation (Original or Offspring Cohort)
According to the World Health Organization (WHO), i.e. rapidly developing signs of focal (or global) disturbance of cerebral function, lasting more than 24 h or leading to death with no apparent cause other than that of vascular origin
2.2 The data source for NAFLD and the selection of instrumental variables
The Million Veteran Program (MVP) consortium started to recruit eligible participants in 2011, and it has become one of the largest biobanks worldwide [
]. The average age of the MVP participants at the time of recruitment was about 64 years. NAFLD was defined by combining the ALT-based approach with non-invasive clinical parameters [
M Vujkovic S Ramdas KM Lorenz , et al. A trans-ancestry genome-wide association study of unexplained chronic alt elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. medRxiv. 20212020.12.26.20248491. doi:10.1101/2020.12.26.20248 491.
], and more details of the definition are listed in Table 1. A multi-ancestry GWAS study was conducted in the MVP consortium comprising 90,408 NAFLD cases and 128,187 controls, identifying 77 SNPs associated with NAFLD [
M Vujkovic S Ramdas KM Lorenz , et al. A trans-ancestry genome-wide association study of unexplained chronic alt elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. medRxiv. 20212020.12.26.20248491. doi:10.1101/2020.12.26.20248 491.
]. Samples were excluded when duplicated, with more heterozygosity than expected, or with ≥2.5% missing genotype calls in the GWAS study. Furthermore, 55 SNPs were identified with European ancestry. To our knowledge, it is the largest scale and latest GWAS study for NAFLD. The original GWAS study was published as a preprint and has not been peer-reviewed yet. The MVP study has been approved by the Central Veterans Affairs Institutional Review Board (IRB) and site-specific Research and Development Committees.
Strict selection criteria were used to select qualified IVs. First of all, we searched for the largest and latest GWAS summary statistics for the genetics proxies of NAFLD. Moreover, 55 SNPs strongly associated with NAFLD in European ancestry were extracted as candidate IVs (P < 5.00E-08). Secondly, the SNPs were eliminated if they were in linkage disequilibrium (r2 < 0.001) or palindromic with intermediate allele frequencies. Thirdly, the SNPs that were not available in the outcome GWAS were excluded, and proxy SNPs were not used in the analysis. Only summary-level GWAS data from European ancestry were used to select the qualified IVs. The F statistics were calculated to present the strength between IVs and NAFLD. Only the SNPs with an F statistic >10 were considered valid and reliable IVs for NAFLD.
Finally, the qualified SNPs were included as IVs to conduct the MR analysis. Detailed information on those IVs is shown in Supplementary Table 1.
2.3 Outcome data
In the current study, we mainly focused on the association of NAFLD with arterial stiffness index, heart failure, coronary heart disease, stroke, ischemic stroke, and its subtypes. We selected the largest published GWAS summary statistic to date for the target outcomes among Europeans. The sources of datasets are shown in Table 1.
Arterial stiffness indicated alterations in arterial hemodynamics [
]. We investigated the causal association between NAFLD and arterial stiffness index (ASI) using the summary statistics obtained from 127,121 UK Biobank participants (age 56 ± 8.1 years, 48.1% males) [
Summary-level data for heart failure were obtained from the Heart Failure Molecular Epidemiology for Therapeutic Targets Consortium (HERMES), comprising 47,309 cases and 930,014 controls of European ancestry [
]. In the large GWAS study for HF, the HERMES consortium comprised of 26 European studies conducted a meta-analysis, and definitions used to adjudicate HF status within each study have been listed in the meta-analysis [
]. The inclusion criteria of HF for this study are listed in Table 1.
For coronary artery disease, the summary statistics were extracted from the Coronary Artery Disease Genome-wide Replication and Meta-analysis (CARDIoGRAM) plus the Coronary Artery Disease (C4D) Genetics Consortium (CARDIoGRAMplusC4D), including 60,801 cases and 123,504 controls [
]. The CARDIoGRAMplusC4D consortium included 48 studies, and more details of the definition of CAD are presented in Table 1.
Stroke and its subtypes were included as outcomes in the study. The full GWAS summary statistic data of stroke were extracted from the MEGASTROKE consortium meta-analysis, including 29 studies. 67,162 any stroke (AS) cases and 60,341 any ischemic stroke (AIS) cases, regardless of subtypes, were included in the original GWAS study [
Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. Toast. Trial of org 10172 in acute stroke treatment.
], the AIS cases were further subtyped into 4373 LAS cases, 7193 CES cases, and 5386 SVS cases. The definitions or diagnostic criteria of the target outcomes were listed in Table 1, and more information was presented in the original GWAS study [
Relevant ethical approvals were obtained from their local institutions, and the ethics statement can be found in each original publication mentioned above.
2.4 Statistical analysis
The two-sample MR analysis was performed to explore the potential causal effects of NAFLD on cardiovascular events, including ASI, HF, CAD, stroke, and ischemic stroke subtypes. As assumed, three assumptions must be satisfied when using the MR method: (1) the selected IVs must be strongly associated with the exposure; (2) the selected IVs should not be associated with potential confounders; (3) the selected IVs could only influence the outcomes through the exposure, but not other pathways (See Supplementary Fig. 1).
In the primary analysis, the conventional random effect inverse variance weighted method (IVW) was used to estimate the causal effect of NAFLD on different CVD events [
]. In addition, five complementary methods were performed as sensitivity analysis, including the weighted median method, the weighted mode method, the MR-Egger regression method, the MR-PRESSO (MR Pleiotropy RESidual Sum and Outlier) [
]. To test and correct the potential horizontal pleiotropy of the selected IVs, the MR-egger regression, MR-PRESSO, and MR-RAPS methods were performed. The MR-egger intercept and zero difference could provide evidence for directional pleiotropy. Additionally, the MR-PRESSO could correct the outliers in the IVs, while the MR-RAPS generated consistent MR results after correcting pleiotropy using robust adjusted profile scores. MR-PRESSO is best suitable if horizontal pleiotropy occurs in less than half of the IVs. MR-RAPS could down weight outliers, and it is sensitive to violations of balanced. A summary of the commonly used MR methods is provided elsewhere by Burgess [
]. The heterogeneity tests were performed to detect the heterogeneity of the IVW method. P < 0.05 in the Cochran's Q test indicated heterogeneity.
Moreover, to replicate the primary results of the MR analysis, we selected two SNPs (rs58542926 on TM6SF2 and rs738409 on PNPLA3) that were strongly associated with NAFLD as IVs to investigate the association of NAFLD with CVDs using IVW method. The two SNPs were regarded as the strongest genetic predictors of NAFLD, which have been used as IVs in the previous MR study [
]. More information on the five IVs is listed in Supplementary Table 2.
Bonferroni correction was applied to avoid false-positive results brought by multiple tests. Therefore, the statistical significance was 0.05 / (1 exposure × 8 outcome) = 0.006. P < 0.05 but above the Bonferroni corrected statistical significance was defined as suggestive evidence for potential causal associations. Additionally, we estimated R2 (the proportion of the IVs that could explain each CVD event), and the statistic power for the MR studies was also calculated using the online website (https://shiny.cnsgenomics.com/mRnd/) [
]. The results and calculation methods are listed in Supplementary Table 3. R (Version 4.0.2) software and TwoSampleMR package were used to perform the MR analysis.
3. Results
We investigated the association of NAFLD with arterial stiffness index, coronary artery disease, heart failure, stroke, ischemic stroke, and its three subtypes (LAS, CES, and SVS). The IVW method provided the primary results. Fig. 1 presents the forest plots of the associations between NAFLD and ASI. The IVW method shows that genetically determined NAFLD was strongly associated with ASI (β = 0.04, 95%CI: 0.02–0.06, P = 5.53E-04). As for HF, the primary results suggested that NAFLD was significantly associated with HF (OR = 1.08, 95%CI: 1.02–1.14 P = 0.005) in the absence of pleiotropy, though the heterogeneity existed (P < 0.001) (Fig. 2). However, there were no significant associations of NAFLD with CAD, AS, AIS, or any AIS subtype in the primary analysis (Figs. 2 and 3).
Fig. 1Mendelian randomization results of NAFLD with arterial stiffness index.
“rs4684847” was removed as outliers in the MR-PRESSO analysis. The GWAS study of arterial stiffness index adjusted for age, age2, sex, weight, genotyping array, device used to obtain pulse waveform, smoking status (current vs non-current smokers), mean arterial pressure and first 10 principal components.
Fig. 2Mendelian randomization results of NAFLD with heart failure and coronary heart disease. “rs11601507”, “rs1169292”, “rs2138157”, “rs2207132”, “rs28929474”, “rs2954038”, “rs429358”, “rs58542926” and “rs738409” were removed as outliers in the MR-PRESSO analysis for coronary artery disease. “rs1169292”, “rs2954038”, and “rs58542926” were removed in the MR-PRESSO analysis for heart failure. The original study of heart failure has adjusted for age, sex, and principal components. And the original study of heart failure has adjusted for Sex, age, and generation (Original or Offspring Cohort).
Fig. 3Mendelian randomization results of NAFLD with stroke, ischemic stroke as well as its subtypes. “rs4841133” was removed as outliers in the MR-PRESSO analysis for ischemic stroke. “rs174535” and “rs738409” were removed in the MR-PRESSO analysis for large artery stroke. The original study of stroke has adjusted for age and sex.
Analyses using the MR-Egger, weighted median, weighted mode, MR-PRESSO, and MR-RAPS were performed to test the robustness of the primary results. As for ASI, robust and consistent estimates were generated without horizontal pleiotropy using five methods (MR-egger intercept: -0.001, P = 0.34) (Fig. 1). After correcting outliers, the MR-PRESSO and MR-RAPS methods observed significant associations of HF and CAD, respectively (Fig. 2). However, the MR-Egger, weighted median, and weighted mode method did not find a significant association of NAFLD with HF or CAD (Fig. 2).
As for stroke, consistent with the primary results, we did not find any significant association of NAFLD with AS, AIS, or any AIS subtype when using the five MR methods (Fig. 3).
3.2 Validation analysis
In the validation analysis using two SNPs strongly associated with NAFLD as IVs, potential evidence suggested that NAFLD was associated with ASI (β = 0.05, 95%CI: 0.01–0.10, P = 0.023). The result remains consistent (β = 0.02, 95%CI: 0.01–0.04, P = 0.001) when five SNPs extracted from an independent GWAS study were used as IVs. However, there was no significant association of NAFLD with other CVD events neither using two SNPs nor five SNPs (P > 0.05) (see Supplementary Figs. 2 and 3).
4. Discussion
To our knowledge, this is the first study leveraging MR to comprehensively investigate the association of genetic predictors determined-NAFLD with CVD events. The study provided genetic support for the causal association between NAFLD and ASI. A suggestive association of NAFLD was hinted at with HF. However, we observed no evidence supporting causal relationships of NAFLD with CAD and any stroke subtype.
Arterial stiffness is a surrogate marker of subclinical atherosclerosis and the surrogate endpoint of CVD [
]. However, it should be noted that arterial stiffness was measured using ASI in the MR analysis. Though ASI is convenient to measure and suitable for large-scale epidemiology studies, it is not a perfect indicator for arterial stiffness, and it may suffer the influence of other hemodynamic properties [
]. Leveraging six complementary MR methods and validation analysis, we found that NAFLD was significantly associated with ASI. However, the mechanism for the effect of NAFLD on atherosclerosis is still uncertain. One possible explanation is that transforming growth factor-β (TGF-β) plays a crucial role in developing fibrosis, and then the TGF-β contributes to the progression of arterial stiffness [
Arterial stiffening with ageing is associated with transforming growth factor-beta1-related changes in adventitial collagen: reversal by aerobic exercise.
]. However, few studies have investigated the causal association between NAFLD and HF. The current study estimated the causal association between NAFLD and HF using the MR design for the first time, and the IVW method suggested that NAFLD might be associated with HF, though the result was not significant in the validation analysis. Even though the underlying mechanism linking NAFLD with HF has not been fully elucidated, NAFLD might lead to myocardial structure subclinical changes, further contributing to the development of HF [
]. Other possible mechanisms might partially account for the association between NAFLD and HF, including insulin resistance, subclinical inflammation, dyslipidemia, et al. [
]. However, a previous MR study reported no causal association between high liver fat content and ischemic heart disease (IHD), though the association was no longer significant using a cohort study design [
]. Of note, only one functional solid variant (PNPLA3 I148M) was used as a genetic instrument for liver fat, and only 633 cases of NAFLD were included in the MR analysis [
], it could not completely substitute for the diagnosis of NAFLD. One genetic variant (such as the PNPLA3 I148M) might not be a good proxy for NAFLD, which might lead to an inconsistent association between NAFLD and CVD [
]. Though the remarkable SNP rs738409 in PNPLA3 was included as one of the selected IV and the largest NAFLD GWAS study was applied in the current study, no significant causal association was found between NAFLD and CAD. More substantial evidence is required to elucidate the role of NAFLD on CAD.
Nevertheless, the associations of NAFLD with stroke seemed to be complicated and inconsistent. A recent meta-analysis comprising 30 studies reported that NAFLD could significantly increase the risk of stroke [
Non-alcoholic fatty liver disease and risk of incident acute myocardial infarction and stroke: findings from matched cohort study of 18 million European adults.
]. We observed that NAFLD was not associated with any stroke type in primary or sensitivity analysis. The MR-RAPS method suggested that NAFLD was significantly associated with AS and AIS (P < 0.05), however, it did not reach the Bonferroni corrected statistical significance threshold. The heterogeneity among different ischemic stroke subtypes and differences in their pathogenesis may also influence the MR results [
Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. Toast. Trial of org 10172 in acute stroke treatment.
]. However, we did not find any significant association between NAFLD and three AIS subtypes. Therefore, the significant associations between NAFLD and stroke found in the observational analysis might be influenced by different confounding factors. Stroke is a kind of complex disease. Genetic factors mainly determined specific stroke subtypes, and the pathogenesis of different subtypes of stroke might be different [
]. Therefore, it still highlights the importance of preventing the risk of stroke in NAFLD patients. Further studies are warranted to clarify the complicated role of NAFLD in stroke and its subtypes.
In general, NAFLD has been discussed as an important cardiovascular risk factor for a long time [
]. Although the study only suggested a robust causal association between NAFLD and ASI, indirect associations due to underlying metabolic factors and potential mechanisms might still explain the associations between NAFLD and other CVD events found in the studies. Patients with NAFLD are often accompanied by other metabolic symptoms, including obesity, abnormal blood glucose, and lipid levels [
], including plasma lipids, inflammation, impaired fibrinolysis, and the contribution of NAFLD to insulin resistance and atherogenic dyslipidemia. For example, the dilated and inflamed visceral adipose tissue of NAFLD generates various molecules, including free fatty acids, tumor necrosis factor α (TNF-α), and other proinflammatory factors, which might finally cause insulin resistance [
]. Additionally, the activation of the NF-κB pathway in NAFLD plays a crucial role in CVD progression in patients with NAFLD by increasing the expression of the intrahepatic cytokine and transcription of proinflammatory-related genes [
To our knowledge, this is the first two-sample MR study that provides new insight into the causal associations between NAFLD and different CVD events. The application of the MR method could reduce bias from confounding and derive robust causal effect estimates. Additionally, we used several essential methods to thoroughly explore the possibility of pleiotropy in IVs, through which we correct the pleiotropy and thus improve the robustness of the MR analysis. In addition, we included two independent MR analyses to validate the results initially observed and conducted a sensitivity analysis using two strongly NAFLD-associated SNPs as IVs, which enhanced the robustness of the results. Of note, the largest and latest GWAS summary statistics for NAFLD provided more information for the IVs and improved the statistic power for the study.
However, the study had some limitations. First of all, the IVs selected accounted for a relatively small genetic variance of phenotypes (ranging from 8.9% to 12.3%), resulting in potentially weak IVs bias. Additionally, the IVs were extracted from a pre-print publication, which has not been peer-reviewed yet. Secondly, a possible limitation of the study is that we were unable to stratify the analysis by NAFLD severity and other factors such as sex and age. The lack of subgroup data for NAFLD is limited to extending these stratified analyses. Thirdly, we mainly focused on European ancestry, potentially reducing the generalizability of our findings. Finally, we cannot completely avoid the influence of horizontal pleiotropy on our results though we have used different methods to exclude outlier variants.
4.2 Future research
Encouragingly, we found solid evidence that NAFLD was significantly associated with ASI. It hints that improvement of NAFLD patients and preventing the CVD risk in NAFLD are necessary. In addition, CVDs and NAFLD might share common cardio-metabolic risk factors, such as obesity and chronic inflammation. Future studies are required to profoundly investigate the natural relationships between NAFLD and CVDs, which might provide new insight into the prevention of CVDs.
5. Conclusion
In conclusion, the MR study supported that NAFLD was casually significantly associated with arterial stiffness index. Our findings do not support a causal relationship between NAFLD and coronary heart disease, heart failure, stroke, and any stroke subtypes. Previous observational associations between NAFLD and CVD events are likely attributed to confounding factors.
CRediT authorship contribution statement
All the authors contributed to the manuscript review and editing. They all approved to submit of the final version of the manuscript.
PHX: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Software; Visualization; Roles/Writing - original draft; Writing - review & editing.
All the authors declare no conflicts of interest in the work.
Acknowledgments
This research was supported by the Special Fund for Health scientific research in public welfare (Grant No. 201502006), the Key Project of National Natural Science Foundation of China (Grant No. 81230066), National Natural Science Foundation of China (Grant No. 81872695), Fujian provincial health technology project (Grant No. 2020CXB009), the Natural Science Foundation of Fujian Province, China (Grant No. 2021J01352), the China Cohort Consortium (Grant No. CCC2020001) and the China Postdoctoral Science Foundation (Grant No. BX2021021). We sincerely acknowledge the contribution from the MVP Consortium, HERMES Consortium, CARDIoGRAMplusC4D Consortium, the MEGASTROKE GWAS dataset, and International Stroke Genetics Consortium, and all concerned investigators and consortiums for sharing the GWAS summary statistics on the diseases.
Non-alcoholic fatty liver disease and risk of incident acute myocardial infarction and stroke: findings from matched cohort study of 18 million European adults.
M Vujkovic S Ramdas KM Lorenz , et al. A trans-ancestry genome-wide association study of unexplained chronic alt elevation as a proxy for nonalcoholic fatty liver disease with histological and radiological validation. medRxiv. 20212020.12.26.20248491. doi:10.1101/2020.12.26.20248 491.
Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. Toast. Trial of org 10172 in acute stroke treatment.
Arterial stiffening with ageing is associated with transforming growth factor-beta1-related changes in adventitial collagen: reversal by aerobic exercise.
00098-1&id=gr1.jpg){kind=link}
00098-1&id=gr2.jpg){kind=link}
00098-1&id=gr3.jpg){kind=link}