Article Text
PDF
Abstract
Aims Previous studies have suggested an association between retinal artery occlusion (RAO) and ischaemic stroke (IS), both associated with elevated inflammatory factors. However, the role of high-sensitivity C-reactive protein (hs-CRP) in the sequential onset of these two diseases is still unclear. Based on this evidence, we evaluated the association of hs-CRP with RAO and IS.
Methods We examined hs-CRP from both the large multicentre cohort study UK Biobank and Chinese Retinal Artery Occlusion study. Cox proportional hazard models were used to study the association of hs-CRP with incident RAO and IS during the long-term follow-up in the UK Biobank. Logistic regression analysis was employed to assess the cross-sectional relationship between hs-CRP with RAO and IS in the Chinese cohort. A restricted cubic spline (RCS) approach was employed to evaluate potential non-linear associations of hs-CRP with IS.
Results After exclusions, the analysis included 459 188 participants from the UK and 338 participants from China. Over a median follow-up of 12.2 years, 136 cases of incident RAO and 3206 cases of incident IS events were recorded in the UK Biobank. After multivariable adjustment, higher hs-CRP (per 10 mg/L) level was associated with increased risks of RAO (HR: 1.34, 95% CI: 1.01 to 1.76) and IS (HR: 1.24, 95% CI: 1.17 to 1.33). RCS analysis revealed a significant non-linear relationship between hs-CRP levels and incident IS (Pnon-linear<0.001). Furthermore, RAO patients with higher hs-CRP levels were more likely to be combined with IS (2.81 mg/L vs 10.14 mg/L, p<0.001). In the Chinese cohort, the association between hs-CRP with RAO and IS was further confirmed. Higher hs-CRP (per 1 mg/L) level was associated with increased risks of RAO (OR: 1.43, 95% CI: 1.15 to 1.78) and IS (OR: 1.13, 95% CI: 1.03 to 1.24).
Conclusions Our findings underscore hs-CRP as a robust risk factor for both RAO and IS. Controlling hs-CRP levels might reduce the incidence of RAO and secondary stroke.
- Ischaemic Stroke
- Epidemiology
- Retina
- Inflammation
Data availability statement
No data are available. This project corresponds to UK Biobank application ID#86091. The UK Biobank data are available on the application to the UK Biobank (www.ukbiobank.ac.uk/).
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
While the association between high-sensitivity C-reactive protein (hs-CRP) and ischaemic stroke (IS) is established, the role of hs-CRP in retinal artery occlusion (RAO) has been limited to single-cohort investigations and remains unclear. Hs-CRP serves as both an acute-phase reactant and a representative marker of inflammation. Elevated levels of circulating hs-CRP have been associated with atherosclerosis and demonstrated predictive and diagnostic value in various types of strokes, especially transient ischaemic attack and IS. Furthermore, there is still a need for research to explore the role of hs-CRP in the sequential onset of RAO and IS.
WHAT THIS STUDY ADDS
This study confirmed that an increase in hs-CRP concentration was associated with higher risks of incident RAO and IS. In addition, RAO patients with higher hs-CRP levels were more likely to be combined with IS.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Our study demonstrated that individuals with elevated hs-CRP levels face a high likelihood of developing RAO and IS. Additionally, our research underscores the link between elevated hs-CRP in RAO patients and incident IS. In community screening efforts, clinicians can implement vigilant monitoring of hs-CRP levels to identify high-risk populations, thus offering tailored prevention and control measures.
Introduction
Retinal artery occlusion (RAO) is an acute ischaemic disease that causes severe visual loss and harbinger of further cardiovascular and cerebrovascular events.1 Ischaemic stroke (IS), another ischaemic disease affecting the brain, is the second leading cause of mortality in people over 60 and the first leading cause of disability.2 3 A previous study highlighted that the retinal vessels have similar embryological, anatomical and physiological characteristics to cerebral vessels in the brain.4–6 Several studies have highlighted a high risk of IS in patients with RAO,7 8 mainly due to shared cardiovascular risk factors such as hypertension, diabetes and dyslipidaemia.9 10 Thus, controlling these common risk factors could offer preventive strategies against RAO and IS, ultimately reducing visual and physical disability risks.
Markers of physiological inflammatory response have been shown to predict stroke in healthy populations. High-sensitivity C-reactive protein (hs-CRP) serves as an acute phase reactant and a representative marker of inflammation, playing a crucial role in the innate immune response.11 Elevated levels of circulating hs-CRP have been linked to atherosclerosis and have shown predictive and diagnostic value in various strokes, particularly transient ischaemic attack (TIA) and IS.12 13 Despite its significance in stroke research, there is limited understanding of the correlation between hs-CRP and RAO.14 Before the onset of RAO, patients often experience vascular transient monocular visual loss, which is akin to a retinal TIA.1 Considering the close connection between RAO, IS and TIA,15 whether hs-CRP acts as a shared risk factor for these conditions and the role of hs-CRP in the sequential onset of these diseases are needed to be understood comprehensively.
In this study, we evaluated the associations between circulating hs-CRP concentration and incident RAO, IS and TIA during the long-term follow-up of UK Biobank (UKB) and assessed whether RAO patients with higher hs-CRP levels were more likely to be combined with IS. In addition, we validated the associations between circulating hs-CRP concentration and incident RAO, IS and TIA in a Chinese cohort.
Materials and methods
Study population
The UKB is a prospective cohort study that recruited over 500 000 participants aged 40–69 years at baseline across 22 assessment centres throughout the UK. Participants provided information on geographical factors, lifestyle and other health-related information through baseline questionnaires, interviews and physical measurements.
The Chinese Retinal Artery Occlusion study consists of participants recruited from the Eye Center of the Second Affiliated Hospital of Zhejiang University (IR2023489). Participants provided information on geographical factors, lifestyle and other health-related information through baseline questionnaires. We excluded individuals who could not be linked to inpatient data.
Ascertainment of hs-CRP at baseline
Blood samples were transferred to a dedicated central laboratory at the end of the collection day and stored at −80°C or in liquid nitrogen until analysis.16 Serum hs-CRP levels were measured by an immunoturbidimetric-high sensitivity analysis on a Beckman Coulter AU5800. Then, we divided hs-CRP concentrations into three groups based on tertiles.
Ascertainment of incident RAO, IS and TIA at follow-up
RAO, IS and TIA cases from both cohorts were ascertained using participants’ self-reported and hospital admissions data, while IS cases from UKB were also identified through linkage to the national death register. Incident RAO was identified through linkage to the International Classification of Diseases (ICD)-10 H340-342. Incident IS was defined as the first occurrence of ICD-10 code I63. ICD-10 codes G45, G458 and G459 were used to identify TIA. The follow-up time was calculated from the date of baseline assessment to the date of incident RAO, IS or TIA events, the date of loss to follow-up, or the end of follow-up (28 April 2021), whichever came first.
IS events were recorded when they occurred around the time of RAO events. The time course of IS and RAO included IS more than 15 days prior, within 15 days prior, simultaneously stroke, within 15 days after, or more than 15 days after RAO.
Assessment of covariates
Covariates pertaining to this study including lifestyle factors, comorbidities and family history were collected during recruitment at baseline. Smoking status was identified through linkage to ‘smoking status’ and recorded as never or former/current cigarette smoking. Alcohol drinker status was identified through linkage to ‘alcohol drinker status’ and recorded as never or former/current alcohol drinking. Diabetes mellitus was identified through linkage to ‘diabetes diagnosed by doctor’, glycated haemoglobin level (HbA1c)≥48 mmol/mol and/or use of insulin and other diabetes-related medication at baseline. Hypertension was identified through self-reported data, systolic blood pressure ≥130 mm Hg or diastolic blood pressure ≥80 mm Hg or anti-hypertensive medications. Hyperlipidaemia was identified through self-reported data, cholesterol ≥6.21 mmol/L or taking stains and other anti-hyperlipidaemia medications.
Socio-demographic factors, including age, sex, ethnicity (recorded as white and others) and education (recorded as graduated or professional school and none) were self-reported. Other covariates included the Townsend deprivation index (an area-based proxy measure for socioeconomic status), physical activity level (recorded as above moderate/vigorous/walking recommendation or not) and body mass index (BMI) which were collected during the initial assessment visit. All the variables from the UKB used in the paper were detailed in online supplemental table 1.
Supplemental material
Statistical analysis
Baseline continuous variables were reported using mean±SD, while categorical variables were summarised as count (percentage). Student’s t-tests were used to test the difference in continuous variables and χ2 test in categorical variables.
Multivariable Cox proportional hazard models were established to assess the association between hs-CRP and incident RAO, IS and TIA. HRs and 95% CIs were estimated. Model 1 was the crude model; model 2 adjusted for age, gender, ethnicity, education, Townsend deprivation index, physical activity, cigarette smoking, alcohol drinking, BMI, family history of stroke, diabetes mellitus, hypertension and hyperlipidaemia. Further, restricted cubic spline (RCS) analysis was applied to test non-linear relationships between the levels of hs-CRP and risks of incident IS and TIA based on multivariable-adjusted Cox proportional hazards models.
In sensitivity analyses, the association between hs-CRP and incident RAO, IS and TIA was examined by excluding individuals who had a prior history of the related brain, cardiovascular diseases and eye diseases (dementia, Parkinson’s disease, neuro-degeneration disorders, coronary heart disease, diabetic-related eye disease, glaucoma, cataract, macular degeneration). Besides, given that acute infections during blood sampling might bias our results, participants with hs-CRP levels more than 10 mg/L were excluded for sensitivity analysis. In addition, taking into consideration the impact of statin therapy on atherosclerosis and inflammation, statin therapy was further adjusted. The hs-CRP levels were compared using Student’s t-tests among patients undergoing RAO alone with those undergoing both RAO and IS, to examine whether hs-CRP is a secondary IS risk factor.
Logistic regression analysis was employed to assess the relationships between hs-CRP with RAO, IS and TIA in the Chinese cohort. ORs and 95% CIs were calculated. The analysis was adjusted for age, gender, BMI, diabetes mellitus and hypertension.
Stata V.17.0 (StataCorp) and R (V.4.2.0, R Foundation for Statistical Computing) were conducted in data analyses. A two-tailed p<0.05 was considered significant.
Results
In the UK cohort, a total of 459 188 individuals with a mean (SD) age of 56.45 (8.09) years and 45.52% men were included in this study, after excluding individuals with a prior history of retinal vascular occlusion, stroke or TIA (figure 1). The range of serum hs-CRP was 0.08–79.96 mg/L (mean (SD): 2.58 (4.34) mg/L). In the Chinese cohort, a total of 338 individuals with a mean (SD) age of 63.27 (11.71) years and 48.82% men were included in this study. The range of serum hs-CRP was 0.6–145.7 mg/L (mean (SD): 5.35 (13.05) mg/L).
Flowchart for population selection from the UK Biobank. CRP, C-reactive protein; TIA, transient ischaemic attack.
Differences in baseline characteristics between UK participants who developed incident outcomes and those who did not develop were described in table 1. Participants with RAO, IS or TIA were more likely to be older, male, former or current smokers, with higher BMI, higher hs-CRP levels, less history of diabetes mellitus and had comorbidities including hypertension and hyperlipidaemia compared with those who did not develop these outcomes. Participants with IS were also more likely to be white ethnicity, less educated, with higher Townsend scores and former or current drinkers compared with those who did not develop IS. Participants with TIA were also more likely to be less educated, with higher Townsend scores and more physical activity compared with those who did not develop TIA.
Baseline characteristics of the study population stratified by the development of incident retinal artery occlusion, ischaemic stroke and transient ischaemic attack
Association of hs-CRP between incident RAO, IS and TIA
Over a median follow-up of 12.2 years, 136 RAO, 3206 IS and 1489 TIA events occurred. Per 10 mg/L increase of hs-CRP concentration at baseline was associated with a 34% higher risk of incident RAO (HR: 1.34, 95% CI: 1.01 to 1.76, p=0.040), 24% higher risk of incident IS (HR: 1.24, 95% CI: 1.17 to 1.33, p<0.001) and 12% higher risk of incident TIA (HR: 1.12, 95% CI: 1.01 to 1.25, p=0.035) after multivariable-adjustment. Participants of the highest hs-CRP group had a higher risk of incident IS (HR: 1.48, 95% CI: 1.33 to 1.64, p<0.001) and TIA (HR: 1.40, 95% CI: 1.20 to 1.64, p<0.001) than the lowest group in the multivariable-adjustment model (table 2). Multivariable-adjusted RCS analyses suggested a non-linear association between hs-CRP levels on a continuous scale and the risk of IS (Pnon-linear<0.001, figure 2A) and TIA (Pnon-linear<0.001, figure 2B).
Association of serum high sensitivity C-reactive protein (hs-CRP) with ischaemic stroke and transient ischaemic attack, allowing for non-linear effects. Association of serum hs-CRP with risks of ischaemic stroke (A), and transient ischaemic attack (B) in the study population. The reference hs-CRP (with HR fixed as 1.0) was 3 mg/L.
Multivariable-adjusted HRs (95% CI) for the association of hs-CRP and incident retinal artery occlusion, ischaemic stroke and transient ischaemic attack
In sex-stratified analyses, hs-CRP (per 10 mg/L) level was significantly associated with IS among both women (HR: 1.22, 95% CI: 1.09 to 1.37, p=0.001) and men (HR: 1.26, 95% CI: 1.16 to 1.35, p<0.001). Hs-CRP (per 10 mg/L) level was significantly associated with TIA among men only (HR: 1.19, 95% CI: 1.05 to 1.35, p=0.007). There was no significant interaction between hs-CRP and sex in the multivariable-adjusted Cox regression model of incident IS and TIA (both Pfor interaction>0.05, (online supplemental table 2). In age-stratified analyses, there was no significant interaction between hs-CRP and age in the multivariable-adjusted Cox regression model of incident RAO, IS and TIA (all Pfor interaction>0.05, (online supplemental table 2). In sensitivity analyses, the association of hs-CRP and incident IS and TIA was consistent with the main findings after excluding individuals with a prior history of related brain, cardiovascular and eye diseases (online supplemental table 3). The association of hs-CRP and incident IS and TIA was consistent with the main findings after excluding participants with hs-CRP levels of more than 10 mg/L (online supplemental table 4). The association of hs-CRP and incident RAO, IS and TIA was consistent with the main findings after further adjusting for statin therapy (online supplemental table 5).
In the Chinese cohort, 18 participants were diagnosed with RAO, 81 with IS and 63 with TIA. After adjusting for age, gender, BMI, diabetes mellitus and hypertension, logistic regression analysis revealed a significant association between hs-CRP (per 1 mg/L) with RAO (OR: 1.43, 95% CI: 1.15 to 1.78, p=0.001) and IS (OR: 1.13, 95% CI: 1.03 to 1.24, p=0.013), but not with TIA (OR: 1.06, 95% CI: 0.97 to 1.16, p=0.225) (online supplemental table 6).
IS events before and after RAO
In the UK cohort, among 136 participants with RAO, 10 patients (7.35%) experienced an IS around a 12.2-year follow-up time. There were four cases of ISs (2.94%) occurring more than 15 days before RAO, three cases of ISs (2.21%) occurring the same day as RAO and three cases of ISs (2.21%) occurring more than 15 days after RAO. Of the patients who had an IS more than 15 days before RAO, these ISs occurred in an average of 2.02 years (ranging from 0.08 years to 3.07 years) before RAO. Of the patients who had an IS more than 15 days after RAO, these ISs cases occurred in an average of 0.97 years (ranging from 0.07 years to 2.71 years) after RAO. RAO patients with higher hs-CRP levels were more likely to be combined with IS (2.81 mg/L vs 10.14 mg/L, p<0.001).
Discussion
In this cross-cohort study, we investigated the association of hs-CRP with incident RAO and IS events during a long-term follow-up. We found that elevated hs-CRP level was significantly associated with increased risks of RAO and IS. The association of hs-CRP between RAO and IS was further confirmed in a Chinese cohort. In addition, RAO patients with higher hs-CRP levels were more likely to be combined with IS.
Over the past decades, studies have been dedicated to understanding the connection between hs-CRP levels with stroke,12 13 IS17 18 and TIA.19 These studies have consistently indicated that higher hs-CRP concentrations were associated with a higher risk of overall stroke.18 20 A meta-analysis of 160 309 people without a history of vascular diseases from 54 long-term prospective studies found that per 1 SD higher log(e) hs-CRP concentration was associated with a 27% increased risk for IS when adjusted for conventional risk factors.21 A prospective study involving 3224 participants demonstrated that high concentrations of hs-CRP were associated with a 28% increased risk of IS.22 A cohort study involving 591 men and 871 women free of stroke/TIA revealed that elevated hs-CRP levels significantly predicted the risk of future IS and TIA in older adults,19 which was consistent with our result. Furthermore, we also found that hs-CRP levels were associated with RAO. Nevertheless, the linkage between hs-CRP and RAO has been sparsely explored, with existing studies suggesting an absence of connection.
Our study revealed a significant association of elevated hs-CRP levels with RAO, IS and TIA within the study population. However, the underlying mechanisms behind these associations remain unclear. Hs-CRP is mainly produced by liver cells under the stimulation of interleukin-6, and is known as an indicator of systemic inflammation.23 24 It has been demonstrated that hs-CRP contributed to different stages of atherogenesis.25 Increased hs-CRP levels marked systemic inflammation and plaque instability,26 leading to plaque rupture and occlusion. Further studies indicated that hs-CRP was also produced by human coronary artery smooth muscle cells in atherosclerotic lesions and directly participated in developing cardiovascular complications.27 28 Moreover, increased hs-CRP levels may contribute to thrombosis by inducing peripheral blood monocytes to synthesise tissue factors.29 Therefore, we believe that hs-CRP may play an essential role in the progression of RAO, IS and TIA. The intricate mechanisms by which hs-CRP influences these conditions warrant further investigation to deepen our understanding of their interconnectedness.
Several studies have demonstrated an increased risk of IS around the time of RAO.10 30 A 12-year cohort study from Korea found that RAO was associated with a 78% increased risk of subsequent stroke after adjusting for comorbidities and socio-demographic factors.31 A population-based study demonstrated that the risk of symptomatic IS was 2.2% in the 15 days before and after a central RAO.32 A Korean study found a stroke rate of 9.18% within 1 year of central RAO and 3.08% within 14 days before and after central RAO.7 In a Mayo Clinic study from 2001 to 2016, 5.3% of patients experienced a symptomatic stroke within 15 days before or after the onset of 300 central RAO cases.33 Our study found that 7.35% of patients had an IS around the time of RAO within 15 days before or after RAO, which was higher than previous studies. The possible explanation was that our research was based on a 12.2-year follow-up time study, while previous studies were almost 1 year or less. RAO can essentially be viewed as a form of acute IS, leading to retinal infarctions.2 The mechanisms and causes closely mirror acute cerebral infarctions in the internal carotid artery territory.1 Increasing evidence of neurovascular complications has prompted experts to consider RAOs as emergencies, necessitating cerebrovascular evaluation within 24 hours.33 Additionally, prompt screening and treatment of vascular risk factors are recommended. Our findings showed that RAO patients with higher hs-CRP levels were more prone to be associated with IS occurrences. Moreover, our research demonstrated that elevated hs-CRP levels were linked to RAO and IS. Thus, hs-CRP emerges as a common serum risk factor for both conditions. Monitoring individuals with elevated hs-CRP levels through accessible tools could prevent RAO and IS effectively. However, further studies are needed to evaluate the predictive ability of elevated hs-CRP levels for RAO combined with IS in a long-term follow-up.
The implications of our findings are highly relevant to public health. Over the past several decades, stroke has become a leading cause of mortality and disability worldwide, carrying substantial economic burdens for post-stroke care. As the population ages, RAO induces severe acute vision impairment and even vision loss in the affected eye among middle-aged and older adults. However, previous studies were limited to using elevated hs-CRP levels to predict RAO and IS. Our study, in contrast, delves deeper by revealing that individuals with elevated hs-CRP levels face a high likelihood of developing RAO and IS. Additionally, our research underscores the link between elevated hs-CRP in RAO patients and incident IS. In community screening efforts, clinicians can implement vigilant monitoring of hs-CRP levels to identify high-risk populations, especially in conjunction with other cardiovascular disease risk factors such as hypertension and hyperlipidaemia. This approach allows for tailored prevention and control measures to be offered to individuals at elevated risk.
Limitations
Several limitations exist for the present study. First, the definitions of RAO, IS and TIA were limited to hospital records, and may have underestimated the incidence as the medical records may fail to identify all cases. Second, the examination of plasma inflammatory markers was a one-time examination and may not be a fasting examination, which may confound the measurements. We supposed that dynamic changes of hs-CRP might be a better biomarker of RAO and IS, but this study is not yet able to evaluate this. Third, causal relationships cannot be established based on our results because of the observational nature of the study.
Conclusion
In summary, our study highlights a significant correlation between elevated hs-CRP levels and the occurrence of RAO and IS. RAO patients with higher hs-CRP levels were more likely to be combined with IS. Controlling hs-CRP levels might reduce the incidence of RAO and secondary stroke.
Data availability statement
No data are available. This project corresponds to UK Biobank application ID#86091. The UK Biobank data are available on the application to the UK Biobank (www.ukbiobank.ac.uk/).
Ethics statements
Patient consent for publication
Ethics approval
Participants from UK Biobank provided written informed consent to investigators of UK Biobank team and the study was approved by the NHS National Research Ethics Service (Ref: 11/NW/0382). This research has been conducted using the UK Biobank Resource under the project number of 86091. The Chinese Retinal Artery Occlusion study was approved by the Second Affiliated Hospital of Zhejiang University Ethics Committee (IR2023489). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
This research used the UK Biobank resource (application number 86091). We thank the participants of UK Biobank for their contribution to the resource.
References
Footnotes
YW, XZ, AL and YN are joint first authors.
Correction notice This paper has been updated since it was first published. The order of the corresponding authors has been changed.
Contributors Study concept and design: HY, LH, XC. Acquisition, analysis or interpretation: All authors. Drafting of the manuscript: YW, XZ. Critical revision of the manuscript for important intellectual content: ZD, FZ, GW, LH, WW, XC, AL, YN, DC, YH, HY. Statistical analysis: YW, XZ. Obtained funding: HY, XZ, AL. Administrative, technical or material support: YF, XS, ZZ, YH, HY, LH, XC. Study supervision: LH, HY, XC. The guarantor is HY.
Funding This study was funded by the National Natural Science Foundation of China (U24A20707, 82301205, 82171075, 82301260), the Natural Science Foundation of Zhejiang Province for Distinguished Young Scholar (R25H120001), the Science and Technology Program of Guangzhou (20220610092), the launch fund of Guangdong Provincial People's Hospital for NSFC (8217040546, 8220040257), Basic and Applied Basic Research Foundation of Guangdong Province (2023B1515120028), Brolucizumab Efficacy and Safety Single-Arm Descriptive Trial in Patients with Persistent Diabetic Macular Edema (BEST) (2024-29), the Medical Scientific Research Foundation of Guangdong Province (A2021378), the China Postdoctoral Science Foundation under Grant (2024T170185). The funders had no role in the study design, data collection, data analysis, data interpretation, or report writing.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.