Introduction
Oxidative stress, a process that occurs when the production of reactive oxygen species (ROSs) exceeds the antioxidant capacity of the body [1], plays a pivotal role in the development of inflammation, hypertension, and atherosclerosis [2,3]. These conditions are key mechanisms underlying the pathogenesis of cardiovascular disease (CVD). Over recent years, the role of oxidative stress in the pathophysiology of CVD has been well established, providing a theoretical foundation for identifying biomarkers of oxidative stress as potential tools for the early detection of CVD [4].
Numerous markers have been developed to evaluate oxidative stress including markers of lipid peroxidation (isoprostane and malondialdehyde) [5,6], and markers of DNA and RNA oxidation (8-oxo-7,8-dihydro-2'-deoxyguanosine and 8-oxo-7,8-dihydroguanosine) [7]. Although these markers have been used in some clinical studies [5-7], the specific measurement of such biomarkers in large population studies is currently challenging due to the technical complexity and cost of the procedures involved [8,9]. Recently, two easy and low-cost assays for oxidative stress biomarkers have been developed: the derivatives of reactive oxygen metabolites (d-ROMs) test and the biological antioxidant potential (BAP) test [10-12]. The d-ROMs test measures the total amount of hydroperoxides, metabolic by-products formed when ROS oxidize essential biomolecules such as lipids, proteins, amino acids, and nucleic acids. This test has been widely utilized in clinical studies to investigate the association between oxidative stress and the prognoses of patients with CVD [13,14], diabetes [15], and colorectal cancer [16]. The BAP test evaluates the capacity of a plasma sample to reduce ferric ions, thus providing a simple, rapid, and cost-effective measurement of antioxidant capacity [17]. The efficiency and practicality of the d-ROMs and BAP tests render then popular tools for clinical research [18-20]. Several studies have investigated associations between d-ROMs, BAP and the risk factors for CVD, including visceral fat and the thickness of the intimamedia in the carotid artery [18,21]. Only one study, utilizing a nested case-control design, reported a positive association between d-ROMs levels and the risk of myocardial infarction (MI) and stroke [22]. These authors found that the levels of total thiol, a marker of redox control status, were inversely associated with the risk of total stroke in healthy European populations [22]. However, the authors did not investigate the association between d-ROMs, thiol levels and different subtypes of stroke. Furthermore, researchers have yet to investigate for an association between BAP levels and the risk of CVD.
In the present study, we investigated the association between plasma markers of oxidative stress and the risk of incident stroke, including different subtypes of stroke (ischemic stroke, subarachnoid hemorrhage, and intraparenchymal hemorrhage), and ischemic heart disease (IHD). This study was conducted in a healthy Japanese population using a case cohort design and employed two oxidative stress biomarkers: d-ROMs and BAP.
Methods
Study population and design
The Japan Public Health Center-based prospective study (JPHC study) is a large-scale prospective study conducted across 11 public health center areas in Japan and consists of two cohorts initiated in 1990 (cohort I) and 1993 (cohort II). At baseline, 140,420 residents aged 40-69 years participated in the survey. Questionnaires on lifestyle and dietary habits were distributed to the participants. In addition, blood samples were acquired and health check-up information was collated. During the follow-up period, the participants were monitored for relocation, vital status, the occurrence of cancer, and CVD (stroke and IHD). Specific details of the JPHC study have been published elsewhere [23].
Following a 5-year interval, a follow-up survey was conducted in 1995 for cohort I and 1998-1999 for cohort II, during which questionnaires, blood samples, and health check-up information were collected. A case-cohort study design was used in this follow-up survey to investigate the associations between d-ROMs, BAP, and incident CVD (Figure 1). Participants were restricted to a base cohort of 29,423 participants (10,672 males and 18,751 females) in nine out of 11 public health center areas, and two metropolitan public health center areas were excluded because information relating to incident CVD was not collected in these areas. In this follow-up survey, participants completed questionnaires and provided samples of venous blood. In the source cohort, 1,805 incident CVD cases (stroke, n=1,497; IHD, n=331) occurred during the follow-up period. After excluding participants with a self-reported history of stroke (n=60), MI (n=45), or angina (n=50); those with occurrences of stroke or IHD before the follow-up survey (n=126); and those without the necessary d-ROMs or BAP data (n=3), 1,521 cases were retained.
We randomly selected 4,946 participants of the source cohort (17% of the source cohort population) as a sub-cohort. After excluding participants with a self-reported history of stroke (n=30), MI (n=25), angina (n=72), stroke, or IHD before the follow-up survey (n=54), as well as those without the necessary d-ROMs or BAP data (n=4), 4,761 participants remained in the sub-cohort. The final case cohort study included 4,761 sub-cohort participants and 1,521 participants with incident stroke or IHD (244 participants were included in the sub-cohort).
Blood collection and the measurement of oxidative stress markers
At the follow-up survey, blood samples (with no fasting requirement) were collected and divided into buffy coat layers and plasma, and then stored at -80°C. In 2023, d-ROMs and BAP were measured in plasma at Kotobiken Medical Laboratories (Tsukuba, Japan). d-ROMs were measured using a reactive oxidative metabolite test kit (Wismerll, Tokyo, Japan) while BAP was measured using a BAP test kit (Wismerll). d-ROMs and BAP levels are expressed in U.CARR and μmol/L, respectively. All laboratory technicians were blinded to information relating to the participants and sub-cohort samples.
Covariate measurements
A self-administered questionnaire was used to collect baseline data, including individual characteristics (e.g., age, sex, height, and weight), lifestyle characteristics (e.g., smoking, drinking, physical activity, and dietary habits), and medical and drug histories.
Prevalent hypertension, hyperlipidemia, or diabetes was identified based on self-reported questionnaire responses indicating a diagnosis by a doctor or the use of anti-hypertensive, anti-diabetic, or anti-hyperlipidemic medications at baseline.
Ascertainment of stroke and coronary heart disease
The occurrence of stroke or IHD was registered in 78 hospitals across nine public health areas. All medical records, including medical imaging data pertaining to stroke and IHD, were reviewed by hospital physicians or physician-epidemiologists.
Stroke was defined according to the criteria of the National Survey of Stroke [24]. All types of stroke, including ischemic stroke, subarachnoid hemorrhage, and intraparenchymal hemorrhage, were confirmed by computed tomography and/or magnetic resonance imaging. Ischemic stroke was further classified into lacunar, large-artery occlusive, and embolic subtypes. Lacunar stroke was defined as the presence of infarcts in the basal ganglia, brainstem, thalamus, internal capsule, or cerebral white matter. Large-artery occlusive stroke was defined as an infarction in cortical regions. Embolic stroke was diagnosed based on the presence of hemorrhagic infarction confirmed by imaging, clinical evidence of cerebral embolism, or medical records indicating a potential embolic source, such as atrial fibrillation, valvular heart disease, or intracardiac thrombus.
IHD included MI and sudden cardiac death. We confirmed the occurrence of MI according to specific criteria (typical chest pain and electrocardiography readings, enzyme findings, or autopsy), as described by the Monitoring Trends and Determinants of Cardiovascular Disease Project [25]. Sudden cardiac death was defined as death from a witnessed cardiac arrest or sudden collapse occurring within one hour of symptom onset without any overt underlying cause.
We used first-ever stroke and IHD events as primary outcomes along with stroke subtypes, including ischemic stroke, subarachnoid hemorrhage, and intraparenchymal hemorrhage. If a participant had experienced both stroke and IHD, both events were included in the analysis. As atherosclerosis underlies both ischemic stroke and IHD, we created a composite outcome for ischemic stroke and IHD to specifically investigate the effects of oxidative stress on these atherosclerosis-related outcomes.
Statistical analysis
The study commenced either with the completion of the questionnaires or with the drawing of blood samples, whichever occurred later. Person-years were calculated from this starting point until the initial occurrence of an incident stroke or IHD, relocation, or the end of follow-up (2009 for cohort I and 2012 for cohort II), whichever occurred first. Descriptive data are presented as numbers and percentages for categorical variables and as means and standard deviations (SDs) for continuous variables. Differences in baseline characteristics between the sub-cohort participants and stroke or IHD cases, as well as across quartiles of d-ROMs and BAP within the sub-cohort, were examined using the t-test or analysis of variance for continuous variables. Differences in proportions were assessed by the chi-squared test. d-ROMs and BAP data, along with d-ROMs/BAP ratios, were categorized into quartiles based on their distribution within the sub-cohort. Hazard ratios (HRs) for total stroke, stroke subtypes, and IHD were calculated for each quartile of d-ROMs, BAP, and d-ROMs/BAP ratio using the weighted Cox proportional hazards method, with the lowest quartile (Q1) as the reference category. To assess linear dose-response relationships, we modeled biomarker levels for every one SD increment in the sub-cohort (d-ROMs: 60 U.CARR, BAP: 266 μmol/L).
Three statistical models were developed, each stratified by area. In Model 1, age and sex were adjusted. In Model 2, additional adjustments were included for body mass index (BMI) (<18.5 kg/m2, 18.5-25 kg/m2, or ≥25 kg/m2), smoking habits (non-smoker or current smoker), alcohol habits (non-regular drinker, regular drinkers of 1-149 or ≥150 g ethanol/week), exercise habits (<1 time/week or ≥1 time/week), walking habits (<3 hours/day or ≥3 hours/day). Considering that an ideal biomarker would be independent of diet [26], we additionally adjusted for green tea (daily drinker or not), coffee (daily drinker or not), and fruit and vegetable consumption (quartiles), the predominant food sources of antioxidants in this cohort [27]. To investigate whether the association between oxidative stress and CVD outcomes was mediated by traditional risk factors for stroke and IHD, we further adjusted for history of hypertension (yes/no), history of diabetes (yes/no), and treatment for hypercholesterolemia (yes/no). This additional adjustment was incorporated into Model 3. The results from Model 2 were considered the primary findings. Missing covariate values were treated as additional missing categories and dummy variables were included in the models.
We assessed Schoenfeld residuals to test the proportional hazards assumption. The residuals for d-ROMs and BAP showed no significant correlation with time, indicating no violation of the proportional hazard assumption.
All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA) and significance was set at a two-tailed probability of 0.05.
Ethical considerations
All participants were informed of the study objectives and completion of the survey questionnaire was considered implicit consent to participate. Blood samples from patients who did not consent were excluded. Before the study began, information was posted on the website of the National Cancer Center, Japan, with the provision for participants to opt out at any time. The study was approved by the institutional review boards of the National Cancer Center (approval no. 2021-206) and University of Tsukuba (approval no. 1907-4).
Results
Compared with the sub-cohort participants, those who developed stroke or IHD tended to be older, male, current smokers, and to consume ≥150 g of ethanol per week (Table 1). Additionally, these participants were more likely to have a history of hypertension. Participants who developed stroke were less likely to walk and consume fruit, had a history of diabetes mellitus, and had higher d-ROMs values. Those who developed IHD were less likely to consume vegetables.
Tables 2 and 3 present sex-specific baseline characteristics of the sub-cohort participants according to d-ROMs and BAP quartiles, respectively. Males and females in the highest d-ROMs quartile tended to be older and to consume ≥150 g of ethanol per week. These participants were also less likely to consume coffee on a daily basis. When considering males, those in the highest quartile had a lower BMI, were more likely to be current smokers, and had a history of hypertension. When considering females, those in the highest quartile had a higher BMI, were more likely to receive treatment for hypercholesterolemia, and were less likely to drink green tea on a daily basis. In contrast, males and females in the highest BAP quartile tended to be younger, consumed more fruit, and were less likely to be current smokers or daily coffee drinkers. They also had a higher BMI. Males in the highest BAP quartile were more likely to receive treatment for hypercholesterolemia and women were more likely to engage in exercise at least once a week.
With a median follow-up of 14.1 years, 1,271 strokes occurred (1,061 outside and 210 within the sub-cohort), including 781 ischemic strokes (647 outside and 134 within the sub-cohort), 146 subarachnoid hemorrhages (128 outside and 18 within the sub-cohort), and 340 intraparenchymal hemorrhages (282 outside and 58 within the sub-cohort). In addition, 265 IHD cases were identified (229 outside and 36 within the sub-cohort).
As shown in Table 4, d-ROMs were positively and linearly associated with the risk of total stroke, ischemic stroke, IHD, and the composite outcome of ischemic stroke and IHD but not with subarachnoid hemorrhage or intraparenchymal hemorrhage (Model 2). Positive associations were consistently observed across subtypes of ischemic stroke (Supplementary Table 1). Further adjustment for potential mediators, including a history of hypertension, diabetes, and treatment for hypercholesterolemia, in Model 3 did not materially alter the associations.
BAP levels were not associated with the risk of CVD outcomes (Table 5 and Supplementary Table 2).
In this study, d-ROMs and BAP were used to evaluate oxidative and antioxidant capacities, respectively. Because oxidative balance is maintained in the body, each measure may be influenced by the other, and simultaneous adjustment for both parameters could obscure true associations. Therefore, d-ROMs and BAP were not included in the same primary model. Notably, including both parameters together did not alter our results (data not shown). In addition, given that the d-ROMs/BAP ratio may reflect the oxidative balance, we conducted additional analysis using the d-ROMs/BAP ratio; these data are shown in Supplementary Table 3. The associations between d-ROMs/BAP ratio and the risk of disease were almost identical to those between d-ROMs and risk; stronger associations were not expected.
Discussion
In this large-scale, population-based cohort study, we identified positive and linear associations between d-ROMs and the risks of total stroke, ischemic stroke, and IHD, but not hemorrhagic stroke. No significant association was found between BAP and the risk of any outcome. This is the first study to comprehensively evaluate the association between oxidative stress biomarkers and CVD risk in the general population.
Our findings are consistent with those of a previous nested case-control study conducted in a European population aged 45-85 years (476 MI cases and 2,380 controls; 454 stroke cases and 2,270 controls) [22]. The authors of this previous study also reported positive associations between d-ROMs levels and the risks of both outcomes, with multivariable HRs per 100 U.CARR increase in d-ROMs of 1.21 (95% CI: 1.05-1.40) for MI and 1.17 (95% CI: 1.01-1.35) for stroke [22].
Our present findings align with established mechanistic evidence that oxidative stress, as quantified by d-ROMs, drives atherosclerosis [28]. Notably, we detected a positive association between d-ROMs levels and both ischemic stroke and IHD, but not with subarachnoid or intraparenchymal hemorrhages. Oxidative stress is known to impair the functionality of vascular endothelial cells, platelets, and erythrocytes, thereby initiating a cascade of events that promote the development of atherosclerosis and thrombus formation [29,30]. In contrast to the occlusive nature of ischemic stroke and IHD, intraparenchymal hemorrhage typically arises from the acute rupture of structurally compromised small arteries [31], while subarachnoid hemorrhage primarily originates from the rupture of intracranial aneurysms [32]. Our present findings suggest that d-ROMs, as a biomarker of oxidative stress, may serve as a specific marker for identifying the risk of occlusive vascular events rather than of hemorrhagic stroke.
No associations were detected between BAP levels and the risk of CVD outcomes. The reason for this lack of association remains uncertain, although we anticipated an inverse association owing to potential antioxidant action. The effective action of most antioxidants depends on interactions between enzymatic and non-enzymatic systems. As the BAP assay primarily reflects non-enzymatic and low-molecular-weight antioxidants, it may not adequately capture antioxidant pathways that are relevant to atherosclerosis [33], potentially explaining the null association observed between BAP levels and the risk of CVD outcomes.
This study has several limitations that should be considered when interpreting our findings. First, although a previous study demonstrated that blood samples for d-ROMs and BAP analysis remained stable when stored at -80°C for up to 60 months [34], the plasma samples used in the present study were stored at -80°C for approximately 20 years. Prolonged storage may have led to autooxidation, potentially affecting the assessment of oxidative and antioxidant capacities. Specifically, autooxidation could have resulted in the underestimation of plasma antioxidant levels. However, a previous cross-sectional study of Japanese men and women with a mean age of 47±11 years reported mean±SD values of 321±63 U.CARR for d-ROMs and 2,829±247 μmol/L for BAP; in the present study, we achieved similar values (333±60 for U.CARR and 2,848±266 μmol/L for BAP). These findings suggest that the influence of long-term storage on auto-oxidation was smaller than expected. Second, we measured d-ROMs and BAP levels at a single time point; this may have led to exposure misclassification during follow-up. Because this potential misclassification is assumed to be non-differential, it would generally exert bias on the effect estimates towards null. Third, because the d-ROMs test reflects global oxidative status, this test cannot distinguish the specific contributions of lipid, protein, or nucleic acid oxidation to overall plasma oxidative stress, thus limiting mechanistic interpretation. Nevertheless, its simplicity and high levels of reproducibility make this test a valuable tool for the early identification of individuals with an elevated risk of CVD. Fourth, because this study was conducted solely in the Japanese population, the generalizability of the findings to other ethnic groups may be limited.
Despite these limitations, the present study had notable strengths. Long-term follow-up allowed the observation of more outcome events, thus enhancing the robustness of the analysis. Furthermore, the distinction between stroke subtypes enabled a more detailed exploration of the influence of oxidative stress on ischemic and hemorrhagic stroke.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
