Available via license: CC BY 4.0
Content may be subject to copyright.
Page 1/26
SexDifference in The Association Between Plasma
Selenium and First Stroke: A Community-Based
Nested Case-Control Study
Huan Hu
the second aliated hospital of nanchang university https://orcid.org/0000-0002-1685-9755
Chonglei Bi
Peoples' Hospital of Rongcheng
Tengfei Lin
China Agricultural University
Lishun Liu
China Agricultural University
Yun Song
China Agricultural University
Binyan Wang
Anhui Medical University
Ping Wang
Sun Yat-Sen University
Ziyi Zhou
China Agricultural University
Chongqian Fang
People's Hospital of Rongcheng
Hai Ma
Health and Family Commission of Rongcheng
Xiao Huang
the second aliated hospital of nanchang university
Lihua Hu
the second aliated hospital of nanchang university
Xiping Xu
China Agricultural University
Hao Zhang
China Agricultural University
Yong Huo
Peking University First Hospital
Xiaobin Wang
Page 2/26
Johns Hopkins University
Huihui Bao
The second aliated hospital of nanchang university
Xiaoshu Cheng
the second aliated hospital of nanchang university
Ping Li ( 183420755@qq.com )
The Second Aliated Hospital of Nanchang University https://orcid.org/0000-0003-3112-0139
Research
Keywords: Selenium, First stroke, First ischemic stroke, First hemorrhagic stroke, Vitamin E
DOI: https://doi.org/10.21203/rs.3.rs-141452/v1
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
Page 3/26
Abstract
Background: To date, there is no clearly dened association between plasma selenium levels and rst
stroke. We aimed to investigate the association between baseline plasma selenium and rst stroke risk in
a community-based, Chinese population.
Methods: Using a nested case-control study design, a total of 1255 rst stroke cases and 1255 matched
controls were analyzed. Participant plasma selenium concentrations were measured by inductively
coupled plasma mass spectrometry (ICP-MS) and the association of plasma selenium with rst stroke
risk was estimated by conditional logistic regression models.
Results: Overall, a nonlinear negative association between plasma selenium with rst total stroke and
rst ischemic stroke risks was found in males, but not in females. Compared with participants with lower
selenium levels (tertile 1-2, <94.1 ng/mL), participants with higher selenium levels (tertile 3, ≥94.1 ng/mL)
had signicantly lower risks of rst total stroke (OR: 0.63; 95% CI: 0.48, 0.83) and rst ischemic stroke
(OR: 0.61; 95% CI: 0.45, 0.83) in males, but not in females with rst total stroke (OR: 0.92; 95% CI: 0.69,
1.22) and rst ischemic stroke (OR: 0.89; 95% CI: 0.65, 1.22). Furthermore, a stronger association between
plasma selenium and rst total stroke was found in males with higher vitamin E levels (≥13.5 μg/mL
vs.
<13.5 μg/mL
P
-interaction=0.007). No signicant association was observed between plasma selenium
and rst hemorrhagic stroke risk in either males or females.
Conclusion: Our study indicated a signicant, nonlinear, negative association between plasma selenium
and rst stroke in males, but not in females.
TRIAL REGISTRATION: ChiCTR1800017274.
Introduction
Stroke is a leading cause of mortality and disability worldwide. Accounting for almost one third of
worldwide stroke mortality, China bears the heaviest stroke burden in the world [1,2]. Since control of risk
factors for stroke helps decrease stroke burden [3,4], the identication of novel risk factors is urgent to
further lower stroke risk. Recently, accumulating evidence has indicated that trace elements might exert
effects on stroke [5,6].
Selenium (Se), an essential trace element, acts as the active center of selenoproteins or selenoenzymes
(eg, glutathione peroxidases), which have many important biological functions including antioxidant,
anti-inammatory and immunoregulatory roles [7-9]. Insucient or excessive selenium intake may be
associated with many adverse health outcomes [10-12]. Particularly, health problems caused by selenium
deciency need to be given great attention in China, a Se-decient country where it is estimated that over
105 million people potentially face adverse health impacts due to selenium deciency [13,14]. Although
cross-sectional epidemiologic studies have indicated inverse associations between selenium levels with
stroke risk [15-17], previous prospective epidemiologic studies have reported inconsistent ndings on the
Page 4/26
associations between selenium concentrations with stroke risk [18-21]. Moreover, few studies have
thoroughly analyzed the potential modiers affecting this association. Therefore, the prospective
relationship between plasma selenium and risk of rst stroke remains inconclusive and deserves further
investigation.
To ll the knowledge gap mentioned above, we performed a nested case-control study to investigate the
association between baseline plasma selenium levels and risk of rst total stroke and stroke subtypes
(ischemic stroke and hemorrhagic strokes), and examined any possible effect modiers using data from
a community-based population in China. To the best of our knowledge, this study is the rst of its kind to
evaluate this relationship in a community-based Chinese population.
Methods
Study population and design
Our present study is a subset of the China H-type Hypertension Registry Study (CHHRS; URL:
http://www.chictr.org.cn; Unique identier: ChiCTR1800017274) which was an ongoing community-based
non-intervention, prospective, observational, multicenter, real-world registry study and was mainly
conducted in Rongcheng county, Shandong province, and Lianyungang, Jiangsu province, China. It was
designed to establish a national registry of patients with hypertension, to investigate the prevalence and
treatment of H-type hypertension in China and the related factors affecting its prognosis, and nally to
construct a risk prediction model of cardio-cerebral and renal vascular diseases. Eligible participants were
men and women aged ≥18 years with essential hypertension, dened as seated systolic blood pressure
(SBP) ≥140 mm Hg and/or seated diastolic blood pressure (DBP) ≥90 mm Hg at the screening visit.
Participants were excluded if they had psychological or nervous system impairment resulting in an
inability to demonstrate informed consent or were unable to be followed-up according to the study
protocol. The trial consisted of 2 stages: screening and recruitment and a 3-year observation follow-up
period. Participants were scheduled for follow-up every 3 months. At each visit, blood pressure, heart rate,
the usage of medications, adverse events, and study outcome events were measured and recorded. The
primary outcome was rst composite of cardiovascular events consisting of nonfatal stroke, myocardial
infarction, and vascular death and all-cause death.
The current nested case-control study utilized data from the CHHRS which had been conducted in
Rongcheng, a coastal area of Shandong province, China. This study matched stroke cases with an equal
number of controls (patients without stroke) by age ± 1 year, sex, and village. Patients with stroke data
from the Chinese Centers for Disease Control and Prevention (CDC, 2016-2018) who had complete
records (physical exam, questionnaire, and biological samples) were selected as cases. The initial sample
consisted of 1401 incident cases and 1401matched controls. Next, we excludedparticipants with
missing serum calcium value (n=287) and unpaired individuals (n=5). Based on the inclusion and
exclusion criteria, 1255 stroke cases and 1255 matched controls with complete calcium measurements
were selected for nal data analysis (Supplementary Figure 1).
Page 5/26
Ethics
The present study was approved by the Ethics Committee of the Institute of Biomedicine, Anhui Medical
University, Hefei, China. All participants signed an approved written consent form after the study protocol
was thoroughly explained to them.
Outcomes
The primary outcome of the present study was a rst nonfatal or fatal stroke. Secondary outcomes
included rst ischemic stroke (fatal and nonfatal), and rst hemorrhagic stroke (fatal and nonfatal).
Information on incidence of rst stroke for all participants was obtained via the Center for Disease
Control and Prevention of Rongcheng county, and checked against the national health insurance system
with electronic linkage to all hospitalizations, or ascertained through active follow-up. Diseases were
coded according to the International Classication of Diseases, 10th Revision (ICD-10). Secondary
outcomes included rst ischemic stroke (I63) and rst hemorrhagic stroke (I60-I61). The primary outcome
(rst nonfatal or fatal stroke) included rst ischemic stroke (I63), rst hemorrhagic stroke (I60-I61) and no
type stroke (I64).
According to government regulations, local authorities from medical institutions are required to report all
new cases of stroke to the local Center for Disease Control and Prevention. A report card which includes
information on demographics, diagnostic basis and date of stroke is required to be submitted on the 28th
of each month. Quality control, including nding and deleting repeated cases, error checking, and
determining any missed cases, is completed by trained ocials. Furthermore, the staff from the local
Center for Disease Control and Prevention would double check these information and also be responsible
for deleting repeated cases and nding logistical errors and missed cases. In addition, 5% of all uploaded
cases are randomly chosen for further conrmation by phone or door-to-door interviews.
Laboratory assays
Baseline serum total homocysteine (tHcy), fasting glucose levels, and lipids were measured using
automatic clinical analyzers (Beckman Coulter, AU680) at the Shenzhen Tailored Medical laboratory in
Shenzhen, China. Estimated glomerular ltration rates (eGFR) were estimated by the Chronic Kidney
Disease Epidemiology Collaboration equation. Baseline plasma vitamin E concentrations were measured
using liquid chromatography–tandem quadrupole mass spectrometry (LC-MS/MS) and plasma selenium
concentrations were measured by inductively coupled plasma mass spectrometry (ICP-MS) using Thermo
Fisher iCAP Q ICP-MS in a commercial laboratory (Beijing DIAN Medical Laboratory, China). In the present
study, the intra-assay CV for selenium ranged from 1.02% to 7.93%, while the inter-assay CV for selenium
ranged from 2.79% to 3.51%. According to a previous study [22], the reference value (50-120 ng/mL) for
plasma selenium levels was used in this study.
Statistical analysis
Page 6/26
Baseline characteristics are presented as means ± SDs for continuous variables and as frequency (%) for
categorical variables. Differences in baseline characteristics between males and females, and cases and
controls, were compared using Chi-square tests for categorical variables and
t-
tests for continuous
variables. Differences in population characteristics according to selenium tertiles were compared using
ANOVA tests, or Chi-square tests, accordingly.
Variables that are known as traditional or suspected risk factors for stroke [23], and matched variables or
variables that showed signicant differences between cases and controls were adjusted for in the
models. Odds ratios (ORs) and 95% condence intervals (95% CIs) for rst stroke, rst ischemic stroke,
and rst hemorrhagic stroke were calculated by modeling plasma selenium as tertiles using conditional
logistic regression, without and with adjustment for matched variables (sex and age), body mass index
(BMI), baseline systolic blood pressure (SBP), baseline diastolic blood pressure (DBP), smoking status,
alcohol consumption, labor intensity, baseline total homocysteine (tHcy), plasma vitamin E, fasting
glucose, estimated glomerular ltration rate (eGFR), antiplatelet drugs, lipoprotein-lowering drugs,
glucose-lowering drugs, antihypertensive drugs, self-reported hypertension, self-reported diabetes, self-
reported atrial brillation, and self-reported hyperlipidemia. A generalized additive model (GAM) and
smooth curve tting (penalized spline method) were evaluated to further characterize the shape of the
association between serum selenium and rst stroke and its subtypes. As additional exploratory
analyses, possible modications of the association between plasma selenium (tertile 3, ≥94.1
vs.
tertile
1-2, <94.1 ng/mL) and rst total stroke in male and female participants were also assessed for variables
including age (<70
vs.
≥70 y), BMI (<24
vs.
≥24kg/m2), current smoking (yes
vs.
no), current alcohol
drinking (yes
vs.
no), baseline SBP (<140
vs.
≥140 mmHg), fasting glucose (<6.1
vs.
≥6.1 mmol/L or
diabetes), totalcholesterol (<5.78 [median]
vs.
≥5.78 mmol/L), triglycerides (<1.17 [median]
vs.
≥1.17
mmol/L), estimated glomerular ltration rate (<90
vs.
≥90 mL/min/1.73m2), total homocysteine (<12.5
[median]
vs.
≥12.5 μmol/L), and vitamin E (<13.5 [median]
vs.
≥13.5 μmol/L) using multivariate logistic
regression models. Diabetes was dened as fasting serum glucose ≥7.0 mmol/L or self-reported use of
anti-diabetic medications, or physician diagnosed diabetes. Potential interactions were examined by
including the interaction terms into those logistic regression models with the greatest number of
confounding variables.
A 2-tailed
P
<0.05 was considered to be statistically signicant in all analyses. R software version 3.4.3
(www.R-project.org) and Empower version 2.17.8 (www.empowerstats.com, X&Y Solutions, Inc.) were
used for all statistical analyses.
Results
Study participants and baseline characteristics
A total of 1255 rst stroke cases (1079 cases of rst ischemic stroke, 171 cases of rst hemorrhagic
stroke, and 5 cases of rst uncertain type of stroke) and 1255 matched controls were included in this
analysis. The mean age of all participants at baseline was 70.75 years (SD, 8.06), 49.48% of the
Page 7/26
participants were male and the mean selenium level was 87.24 ng/mL (SD, 18.57). Baseline
characteristics of male and female participants are shown in Table 1. Detailed plasma selenium
concentration distribution of subjects is listed in Supplemental Table 1 and 95.1% of the participants
were within the normal range of selenium levels as the reference value (50-120 ng/mL) for plasma
selenium concentrations. As shown in Table 1, male participants had non-signicantly higher selenium
levels than females (87.68 ± 18.58
vs.
86.80 ± 18.55 ng/mL;
P
=0.230). Males also tended to be older,
were more likely to be current smokers and current drinkers, had higher DBP and tHcy levels, as well as
lower BMI, SBP, TC, TG, glucose, and vitamin E levels at baseline and a lower frequency of lipid-lowering,
glucose-lowering and antihypertensive drug use, and were less likely to be hypertensive patients, or self-
reported diabetic and self-reported hyperlipidemia patients compared with female participants.
Baseline characteristics of cases and control participants are shown in Table 2. Stroke cases had non-
signicantly lower selenium levels than controls (86.63 ± 17.63
vs.
87.84 ± 19.45 ng/mL;
P
=0.101).
Stroke cases also tended to have higher BMI, baseline BP, TG, fasting glucose, tHcy levels, and a higher
frequency of antiplatelet, glucose-lowering and antihypertensive drug use, were more likely to be current
smokers, hypertensive and self-reported diabetic patients, as well as have lower high-density lipoprotein-
cholesterol levels at baseline compared with controls. In addition, plasma selenium was positively
associated with BMI, current smoking, current alcohol drinking, self-reported diabetes, higher frequency of
glucose-lowering drug use, TC, high-density lipoprotein-cholesterol, fasting glucose, and vitamin E levels
and was inversely associated with labor intensity and tHcy levels at baseline (Supplemental Table 2).
Association between plasma selenium concentration and rst stroke in total participants
Overall, there was a nonlinear negative association between plasma selenium levels with the risk of rst
total stroke and rst ischemic stroke (Fig. 1A and Fig. 1B), but not with the risk of rst hemorrhagic stroke
(Supplemental Fig. 2A) in total participants. Consistently, when plasma selenium was assessed as
tertiles, signicantly lower risks of rst total stroke (Model 2, OR: 0.77; 95%CI: 0.61, 0.97) and rst
ischemic stroke (Model 2, OR: 0.78; 95%CI: 0.60, 0.99) were found in participants in tertile 3 (≥94.1
ng/mL) than in those in tertile 1 (<79.1 ng/mL) (Table 3). Due to the similar rst total stroke and rst
ischemic stroke prevalence in participants with selenium levels in tertile 1 and tertile 2 (Table 3), we
combined these two groups into one group called tertile 1-2. Compared with participants with lower
selenium levels in tertile 1-2 (<94.1 ng/mL), signicantly lower risks of rst total stroke (Model 2, OR: 0.76;
95%CI: 0.63, 0.93) and rst ishcemic stroke (Model 2, OR: 0.75; 95%CI: 0.61, 0.93) were found in those
with a higher selenium levels in tertile 3 (≥94.1 ng/mL) (Table 3). However, no signicant association
was found between plasma selenium concentrations and rst hemorrhagic stroke (Table 3).
Association between plasma selenium concentrations and rst stroke by sex
Given the differences in plasma selenium levels between male and female participants (87.68 ± 18.58
vs.
86.80 ± 18.55 ng/mL), we further investigated the possible effect of sex on the selenium-rst stroke
association. Overall, there was a nonlinear negative association between plasma selenium levels with the
risks of rst total stroke and rst ischemic stroke in males (Fig. 1C and Fig. 1D), but not in females (Fig.
Page 8/26
1E and Fig. 1F). Furthermore, there was no signicant association between plasma selenium and rst
hemorrhagic stroke in both sexes (Supplemental Fig. 2B-C). Consistently, when plasma selenium was
assessed as tertiles, the highest tertile (T3, ≥94.1 ng/mL) of plasma selenium was associated with a
lower rst total stroke risk in males (Model 2, OR: 0.67; 95%CI: 0.48, 0.93,
P=
0.017), but not in females
(Model 2, OR: 0.85; 95%CI: 0.61, 1.19,
P=
0.353) compared with the lowest tertile (T1, <79.1 ng/mL) of
plasma selenium (Table 4). Accordingly, higher selenium levels in tertile 3 (≥94.1 ng/mL) were
associated with a lower rst total stroke risk in males (Model 2 OR: 0.63; 95%CI: 0.48, 0.83,
P=
0.001), but
not in females (Model 2, OR: 0.92; 95%CI: 0.69, 1.22,
P=
0.563) compared with lower selenium levels in
tertile 1-2 (<94.1 ng/mL) (Table 4).
Similar effects of sex on the selenium-rst ischemic stroke association were also observed and are
displayed in Table 4. However, no signicant association was found between plasma selenium and rst
hemorrhagic stroke risk among both males and females (Table 4).
Stratied analysis by potential effect modiers in male and female participants
Stratied analyses were conducted to explore potential modiers affecting the association between
plasma selenium (tertile 3, ≥94.1
vs.
tertile 1-2, <94.1 ng/mL) and rst total stroke risk among male
participants (Table 5). A stronger nonlinear negative association between baseline plasma selenium and
rst total stroke was found among males with higher vitamin E levels compared to those with lower
vitamin E levels (≥13.5 μg/mL; OR, 0.39; 95%CI: 0.25, 0.60;
vs.
<13.5 μg/mL; OR, 0.85; 95%CI: 0.62, 1.17;
P
for interaction=0.007). None of the other variables, including age (<70
vs.
≥70 years), BMI (<24
vs.
≥24
kg/m2), current smoking (yes
vs.
no), current alcohol drinking (yes
vs.
no), baseline SBP (<140
vs.
≥140
mmHg), fasting glucose (<6.1
vs.
≥6.1 mmol/L or diabetes), TC (<5.78 [median]
vs.
≥5.78 mmol/L), TG
(<1.17 [median]
vs.
≥1.17 mmol/L), eGFR (<90
vs.
≥90 mL/min/1.73m2), and total homocysteine (<12.5
[median]
vs.
≥12.5 μmol/L) were found to modify the association between plasma selenium (tertile 3,
≥94.1
vs.
tertile 1-2, <94.1 ng/mL) and the risk of rst stroke in males (
P
for all interactions >0.05).
Furthermore, none of the above variables signicantly modied the association of plasma selenium and
the risk of rst total stroke in female participants (
P
for all interactions >0.05) (Supplemental Table3).
Discussion
This nested case-control study demonstrates that higher baseline plasma selenium is associated with
lower risks of rst total stroke and rst ischemic stroke in males, but not in females. Plasma vitamin E
levels signicantly modied the association between plasma selenium and rst total stroke in males.
Furthermore, no signicant association was found between plasma selenium and rst hemorrhagic
stroke risk in either male or female participants.
Conflicting ndings of the association between plasma selenium and the risk of stroke have been
reported by previous studies. A nested case-control study [20] enrolling 1304 stroke cases found that
higher plasma selenium levels were signicantly associated with a lower risk of hemorrhagic stroke, but
Page 9/26
not ischemic stroke; the odds ratios (ORs) of hemorrhagic and ischemic stroke were 0.68 (95%CI: 0.51,
0.91) and 0.92 (95%CI: 0.82, 1.05) in the higher selenium levels in tertile 3 (compared with tertile 1). One
case-control study [17] including 1277 ischemic stroke patients indicated that higher plasma selenium
levels were associated with a decreased risk of ischemic stroke, where the OR for those with higher
selenium levels in quartile 4 (compared with quartile 1) was 0.10 (95%CI: 0.06, 0.17). Moreover, the
Canadian Health Measures Survey (CHMS) and the National Health and Nutrition Examination Study
(NHANES) found inverse, cross-sectional associations between whole blood selenium and prevalence of
stroke, as well as the Inuit Health Survey (IHS) indicated a reverse relation of whole blood and dietary
selenium levels with stroke, but revealed an L-shaped relationship [15,16]. However, the Reasons for
Geographic and Racial Differences in Stroke Study (REGARDS) [18] revealed that higher environmental
selenium levels were associated with increased stroke risk; the hazard ratio (HR) for those with higher
selenium levels in quartile 4 (compared with quartile 1) was 1.33 (95%CI: 1.09, 1.62). It is noteworthy that
all of these studies used different sources (plasma, whole blood, diet, and environment) of selenium
levels to assess the association between selenium levels and stroke, which might be one reason for the
discrepancy of these ndings.
Several studies have also explored the association between selenium and stroke mortality, specically. A
cohort study [19] enrolling 23 stroke death cases among 1100 Finnish males found that low serum
selenium (<45μg/L) was associated with a higher risk of stroke mortality, reporting an adjusted relative
risk of 3.7 (95%CI: 1.0, 13.1). The NHANES III cohort study [24] including 13887 participants found that
the association curve for selenium and stroke mortality had a reversed U-shape. However, another cohort
study including 1103 Chinese participants found no signicant association of plasma selenium levels
and stroke mortality [25]. Notably, these studies focused on stroke mortality and these ndings might not
represent the association of plasma selenium levels and rst stroke risk. Similarity, prospective
associations between selenium status/intake and cardiovascular outcomes remain inconclusive [26-28].
None of the above research reported a sex difference in the association between plasma selenium and
stroke risk, and the results of these studies remain inconclusive. The present study provides an
opportunity to explore the possible relation of plasma selenium and rst stroke, and to examine the
potential effect modiers in a community-based Chinese population.
Our current study provides three new insights into the eld. First, to the best of our knowledge, this is the
rst study to nd a signicantly nonlinear, inverse association between plasma selenium with rst total
stroke and rst ischemic stroke risks in males, but not in females. The differences in the primary outcome
(rst stroke) between sex in our study may be explained by the differences in the way selenium is
metabolized between the male and female reproductive systems. The retention rate for selenium is highly
ecient in the testes, while it appears that the female reproductive system does not retain signicant
levels of selenium as eciently [29-31]. In addition, the testes might compete for selenium utilization with
the brain under selenium-compromised conditions [32], suggesting that the brain in males might be more
susceptible to selenium deciency than in females. The interaction of selenium with the thyroid axis may
be another reason for the differences. Wang et al. [33] has demonstrated strong sex-specic differences
in risk and development for hyperthyroidism in relation to baseline selenium intake, selenium deciency
Page 10/26
might constitute a risk factor for hyperthyroidism in males but no substantial association was found
between hyperthyroidism prevalence and selenium status in females. Since hyperthyroidism has been
reported to associate with 2- to 3-fold increased risk for ischemic and non-ischemic stroke [34]. Therefore,
we speculate that high selenium levels may reduce the adverse effects in males due to the selenium
deciency, which may explain why the nonlinear inverse association between serum selenium and rst
stroke was mainly found among males. Further prospective studies are needed to verify this differential
association by sex.
Second, we observed a sharp decline in the risk of rst stroke when plasma selenium was over 94.1
ng/mL, suggesting that this value might serve as a high plasma selenium cut-off point marking a
decreased risk of rst stroke or a low plasma selenium cut-off point marking an increased risk of rst
stroke. This cut-off value agrees with a previous study which reported that plasma selenium >90μg/L was
sucient to optimize functions of selenoproteins [35], which are believed to carry out the functions of
selenium in the role of selenium compounds. Schomburg Lutz et.al [36] also reported that deciency of
Selenoprotein-P, the main carrier of selenium to target organs and reduces tissue oxidative stress both
directly and by delivering selenium to protective selenoproteins, was associated with increased risk of
stroke in a North European population without history of cardiovascular disease. However, it should also
be noted that this cut-off value is still within the normal range for human plasma selenium (50-120
ng/mL) [22], and our ndings were found mainly among a population with normal selenium levels, with a
prevalence of plasma selenium <50, 50 to 120, and >120 ng/mL of 1.2%, 95.1%, and 3.7% in this study
(Supplemental Table 1). Therefore, the use of the cut-off value for plasma selenium concentration among
stroke patients needs careful consideration. Our results, if further conrmed, might have vital clinical and
public health implications for community residents in China.
Third, our study is the rst to indicate a stronger nonlinear negative relation of plasma selenium and rst
stroke in male participants with higher plasma vitamin E levels (≥13.5 μg/mL) than those with lower
plasma vitamin E levels (<13.5 μg/mL). This nding suggests that higher plasma selenium and higher
plasma vitamin E levels may jointly decrease the rst stroke risk. A previous meta-analysis demonstrated
a signicant inverse association between dietary vitamin E intake and stroke risk, where a higher dietary
vitamin E intake was associated with a lower risk of stroke [37]. The exact mechanisms underlying a high
selenium × high vitamin E interaction remain unclear. One plausible biological explanation for the
interaction may be due to that both selenium and vitamin E belong to vital antioxidants and participate in
protecting against brain oxidative stress [38], one of the hallmarks of stroke. Accordingly, high plasma
vitamin E and selenium levels may share some cellular and molecular mechanisms for the pathogenesis
of stroke, which could cause the interaction in the nonlinear negative relation of plasma selenium and
rst stroke in males. Further studies are warranted to verify this hypothesis.
While the mechanisms underlying the effect of selenium on rst stroke remain inconclusive, an
association seems reasonable due to several vital biological functions of selenium. Hosnedlova et al. [39]
demonstrated that selenium mainly exerts a protective effect against oxidative lipid damage of the brain
and modulates neurotoxicity and oxidative stress in the nervous tissue. Furthermore, modulation of
Page 11/26
inammatory and metabolic signaling, as well as preservation of mitochondrial function may also be
involved in the protective role of selenium on stroke [40,41]. Selenium deciency in heart failure patients
was independently associated with impaired exercise tolerance and a 50% higher mortality rate, and
impaired mitochondrial function in vitro, in human cardiomyocytes [42]. Ishrap et al. [43] reported that
pharmacological selenium supplementation might have an unexpected ability to drive adaptive
transcription to counter ferroptosis and protect neurons after stroke both in vitro and in vivo in animal
models. Further studies are needed to illuminate the mechanisms underlying the association between
plasma selenium and stroke.
Several possible limitations in this study should be mentioned. First, the plasma selenium concentrations
only represent the baseline selenium levels of all participants; more frequent measurements during the
follow-up would have strengthened the accuracy of our results. Second, only plasma selenium
concentrations were used as the biomarker of selenium levels in our study; other biomarkers including
whole blood and urinary selenium concentrations should also be considered when performing a
sensitivity analysis to conrm our ndings. Third, this was a nested, case-control study with a relatively
small sample size from a community-based population and all stratied analyses were not prespecied,
thus, this work was a product of hypothesis generating, and further larger-scale cohort studies are needed
to verify the ndings. Finally, since selenium is renally eliminated under the inuence of diuretics, we
adjusted all antihypertensive drugs together and did not analyze the effects of diuretic separately on the
association, thus, further analysis is need to clarify this issue.
Perspectives and signicance
In summary, we found a signicantly nonlinear, inverse association between baseline plasma selenium
and the risks of rst stroke and rst ischemic stroke in males, but not in females. In addition, no
signicant association between plasma selenium and rst hemorrhagic stroke was found among either
sex. If further conrmed, our ndings may provide useful data for clinical and nutritional guidelines on
the primary prevention of rst stroke, by taking plasma selenium into account to serve as a potentially
modiable risk factor, and a possible biomarker for purposes of monitoring and intervention.
Abbreviations
CHHRS, China H-type Hypertension Registry Study; CDC, Chinese Centers for Disease Control and
Prevention; OR, odds ratio; CI, condence intervals; BMI, body mass index; SBP, systolic blood pressure;
DBP, diastolic blood pressure; tHcy, total homocysteine; eGFR, estimated glomerular ltration rate; TC,
total cholesterol; TG, triglycerides; HDL-C, high-density lipoprotein-cholesterol.
Declarations
Acknowledgements
Page 12/26
We acknowledge the contribution of all staff who participated in the present study as well as the study
participants who shared their time with us.
Authors’ contributions
Concept and design of this study: Dr. Ping Li and Xiping Xu; Manuscript composition: Dr. Huan Hu, and
Ping Li; Data acquisition and collation: Lishun Liu; Statistical analysis: Huan Hu, Ping Wang and Ziyi
Zhou; Reviewed and revised the manuscript: Ping Li, Xiao Huang, Huihui, Bao, and Xiping Xu. The other
authors coordinated this analysis. All authors read and approved the nal manuscript.
Funding
The study was supported by funding from the following: the National Key Research and Development
Program [2016YFE0205400, 2018ZX09739010, 2018ZX09301034003], the Science and Technology
Planning Project of Guangzhou, China [201707020010]; the Science, Technology and Innovation
Committee of Shenzhen [GJHS20170314114526143, JSGG20180703155802047]; the Economic, Trade
and Information Commission of Shenzhen Municipality [20170505161556110, 20170505160926390];
the National Natural Science Foundation of China [81960074, 81860058, 81500233, 81560079]; the
Jiangxi Outstanding Person Foundation [20192BCBL23024], the Major Projects of the Science and
Technology Department, Jiangxi [20171BAB205008, 20152ACB20022], the Funding Scheme for
Academic and Technical Leaders of Major Disciplines, Jiangxi [20172BCB22027], and Special Funds for
Guiding Local Scientic and Technological Development by the Central Government of China
(S2019CXSFG0016).
Conict of Interest Disclosures:
Dr. Xiping Xu reports grants from the National Key Research and Development Program
[2016YFE0205400, 2018ZX09739010, 2018ZX09301034003), the Science and Technology Planning
Project of Guangzhou, China (201707020010), the Science, Technology and Innovation Committee of
Shenzhen [GJHS20170314114526143, JSGG20180703155802047), and the Economic, Trade and
Information Commission of Shenzhen Municipality [20170505161556110, 20170505160926390].
Dr. Xiao Huang reports grants from the National Natural Science Foundation of China [81960074,
81500233], the Jiangxi Outstanding Person Foundation [20192BCBL23024], and Major projects of the
Science and Technology Department, Jiangxi [20171BAB205008].
Dr. Ping Li reports grants from the National Natural Science Foundation of China [81560079, 81860058],
Major Projects of the Science and Technology Department, Jiangxi, [20152ACB20022], Funding Scheme
for Academic and Technical Leaders of Major Disciplines, Jiangxi [20172BCB22027], and Special Funds
for Guiding Local Scientic and Technological Development by the Central Government of China
(S2019CXSFG0016). No other disclosures were reported.
Availability of data and materials
Page 13/26
All data are available from the corresponding author upon request.
Ethics approval and consent to participate
The protocol of the present study was approved by the the Ethics Committee of the Institute of
Biomedicine, Anhui Medical University, Hefei, China. All participants signed an approved written consent
form.
Competing interests
The authors declare that they have no competing interests.
Author details
1Department of Cardiovascular Medicine, the Second Aliated Hospital of Nanchang University, No. 1
Minde Road, Nanchang, Jiangxi Province, China. 2Center for Prevention and Treatment of Cardiovascular
Diseases, the Second Aliated Hospital of Nanchang University, Nanchang, No. 1 Minde Road,
Nanchang, Jiangxi Province, China. 3People's Hospital of Rongcheng, No. 298 Chengshan Avenue,
Rongcheng, Shandong Province, China. 4Beijing Advanced Innovation Center for Food Nutrition and
Human Health, College of Food Science and Nutritional Engineering, China Agricultural University, No. 17
Tsinghua East Road, Beijing, China. 5Institute of Biomedicine, Anhui Medical University, No. 81 Meishan
Road, Hefei, Anhui Province, China. 6Shenzhen Evergreen Medical Institute, No. 16 Gaoxin Middle 1 Road,
Shenzhen, China. 7School of Public Health (Shenzhen), Sun Yat-Sen University, No. 135 Xingang West
Road, Guangzhou, Guangdong Province, China. 8Health and Family Planning Commission, No. 688
Qingshan East Road, Rongcheng, Shandong Province, China. 9 Department of Cardiology, Peking
University First Hospital, No. 8 Xishiku Street, Beijing, China. 10Department of Population, Family and
Reproductive Health, Johns Hopkins University Bloomberg School of Public Health, 3400 N. Charles
Street, Baltimore, MD, 21205, USA
References
1. Zhou M, Wang H, Zeng X, Yin P, Zhu J, Chen W, et al. Mortality, morbidity, and risk factors in China
and its provinces, 1990-2017: a systematic analysis for the Global Burden of Disease Study Lancet.
2019; 394: 1145-1158.
2. Wang W, Jiang B, Sun H, RuX, SunD, WangL, et al. Prevalence, Incidence, and Mortality of Stroke in
China: Results from a Nationwide Population-Based Survey of 480 687 Adults. Circulation. 2017;
135: 759-771.
3. Wu S, Wu B, Liu M, ChenZ, WangW, AndersonCS, et al. Stroke in China: advances and challenges in
epidemiology, prevention, and management. Lancet Neurol. 2019; 18: 394-405.
4. Lackland D.T, Roccella E.J, Deutsch A.F, FornageM, GeorgeM.G, HowardG, et al. Factors inuencing
the decline in stroke mortality: a statement from the American Heart Association/American Stroke
Page 14/26
Association. Stroke. 2014; 45: 315-353.
5. Uesugi S, Ishihara J, Iso H, SawadaN, TakachiR, InoueM, et al. Dietary intake of antioxidant
vitamins and risk of stroke: the Japan Public Health Center-based Prospective Study. Eur J Clin Nutr.
2017; 71: 1179-1185.
6. Mohammadifard N, Humphries KH, Gotay C, Mena-SánchezG, Salas-SalvadóJ, Esmaillzadeh.A, et
al. Trace minerals intake: Risks and benets for cardiovascular health. Crit Rev Food Sci Nutr. 2019;
59: 1334-1346.
7. Rayman M.P. Selenium and human health. Lancet. 2012; 379: 1256-1268.
8. Roman M, Jitaru P, Barbante C. Selenium biochemistry and its role for human health. Metallomics.
2014; 6: 25-54.
9. Zoidis E, Seremelis I, Kontopoulos N, DanezisGP. Selenium-Dependent Antioxidant Enzymes: Actions
and Properties of Selenoproteins. Antioxidants (Basel). 2018; 7: 66.
10. Loscalzo J. Keshan disease, selenium deciency, and the selenoproteome. N Engl J Med. 2014; 370:
1756-1760.
11. Fang LQ, Goeijenbier M, Zuo SQ, WangLP, LiangS, KleinSL, et al. The association between
hantavirus infection and selenium deciency in mainland China. Viruses. 2015; 7: 333-351.
12. Wang XL, Yang TB, Wei J, LeiGH,ZengC. Association between serum selenium level and type 2
diabetes mellitus: a non-linear dose-response meta-analysis of observational studies. Nutr J. 2016;
15: 48-56.
13. Xu ZC, Shao HF, Li S, Zheng C. Relationships between the selenium content in ue-cured tobacco
leaves and the selenium content in soil in ENSHI, China tobacco-growing area. Pak J Bot. 2012; 44:
1563-1568.
14. Dinh QT, Cui Z, Huang J, TranTAT, WangD, YangW, et al. Selenium distribution in the Chinese
environment and its relationship with human health: A review. Environ Int. 2018; 112: 294-309.
15. Hu XF, Stranges S, Chan LHM. Circulating Selenium Concentration Is Inversely Associated With the
Prevalence of Stroke: Results From the Canadian Health Measures Survey and the National Health
and Nutrition Examination Survey. J Am Heart Assoc. 2019; 8: e012290.
16. Hu XF, Sharin T, Chan HM. Dietary and blood selenium are inversely associated with the prevalence
of stroke among Inuit in Canada. Trace. Elem. Med. Biol. 2017; 44: 322-330.
17. Wen Y, Huang S, Zhang Y, ZhangH,ZhouL,LiD, et al. Associations of multiple plasma metals with
the risk of ischemic stroke: A case-control study. Int. 2019; 125: 125-134.
18. Merrill PD, Ampah SB, He K, RembertNJ, BrockmanJ, KleindorferD, et al. Association between trace
elements in the environment and stroke risk: The reasons for geographic and racial differences in
stroke (REGARDS) study. J Trace Elem Med Biol. 2017; 42: 45-49.
19. Virtamo J, Valkeila E, Alfthan G, PunsarS, HuttunenJK, KarvonenMJ. Serum selenium and the risk
of coronary heart disease and stroke. Am J Epidemiol. 1985; 122: 276-282.
Page 15/26
20. Xiao Y, Yuan Y, Liu Y, YuY, JiaN, ZhouL, et al. Circulating Multiple Metals and Incident Stroke in
Chinese Adults. Stroke. 2019; 50: 1661-1668.
21. Jenkins DJA, Spence JD, Giovannucci EL,KimYI, JosseR, ViethR, et al. Supplemental Vitamins and
Minerals forCVD Prevention and Treatment. J Am Coll Cardiol. 2018; 71: 2570-2584.
22. Wilhelm M, Ewers U, Schulz C. Revised and new reference values for some trace elements in blood
and urine for human biomonitoring in environmental medicine. Int J Hyg Environ Health. 2004; 207:
69-73.
23. Meschia JF, Bushnell C, Boden-Albala B, BraunLT, BravataDM, ChaturvediS, et al. Guidelines for the
primary prevention of stroke: a statement for healthcare professionals from the American Heart
Association/American Stroke Association. Stroke. 2014; 45: 3754-3832.
24. Bleys J, Navas-Acien A, Guallar E. Serum selenium levels and all-cause, cancer, and cardiovascular
mortality among US adults. Arch Intern Med. 2008; 168: 404-410.
25. Wei WQ, Abnet CC, Qiao YL, DawseySM, DongZW, SunXD, et al. Prospective study of serum
selenium concentrations and esophageal and gastric cardia cancer, heart disease, stroke, and total
death. Am J Clin Nutr. 2004; 79: 80-85.
26. Stranges S, Navas-Acien A, Rayman MP, GuallarE. Selenium status and cardiometabolic health: state
of the evidence. Nutr Metab Cardiovasc Dis. 2010; 20: 754-760.
27. KhanSU, KhanMU, RiazH, ValavoorS, ZhaoD, VaughanL, et al. Effects of Nutritional Supplements
and Dietary Interventions on Cardiovascular Outcomes: An Umbrella Review and Evidence Map. Ann
Intern Med.2019; 171: 190-198.
28. ReesK, HartleyL, DayC, FlowersN, ClarkeA, StrangesS. Selenium supplementation for the primary
prevention of cardiovascular disease. Cochrane. Database Syst Rev. 2013; 31: CD009671.
29. Brown DG, Burk RF. Selenium retention in tissues and sperm of rats fed a Torula yeast diet. J Nutr.
1973; 103: 102-108.
30. G Hardy, I Hardy. Selenium: the Se-XY nutraceutical. Nutrition. 2004; 20: 590-593.
31. L.Hateld, U.Schweizer, P.A.Tsuji. V.N.Gladyshev. L Schomburg. Sex-specic differences in biological
effects and metabolism of selenium, in: D L Hateld, et al. 4th ed. Selenium - Its Molecular Biology
and Role in Human Health, Springer, New York, NY, 2016. pp. 214-220.
32. PittsMW, KremerPM, HashimotoAC, TorresDJ, ByrnsCN, WilliamsCS, et al. Competition between
the Brain and Testes under Selenium-Compromised Conditions: Insight into Sex Differences in
Selenium Metabolism and Risk of Neurodevelopmental Disease. J Neurosci. 2015; 35: 15326-15338.
33. WangY,ZhaoF,RijntjesE,WuL,WuQ,SuiJ, et al. Role of Selenium Intake for Risk and
Development of Hyperthyroidism. J Clin Endocrinol Metab. 2019; 104: 568-580.
34. Dekkers OlafM, Horváth-Puhó Erzsébet, Cannegieter SuzanneC, Vandenbroucke JanP, Sørensen
HenrikToft, Jørgensen JensOttoL. Acute cardiovascular events and all-cause mortality in patients
with hyperthyroidism: a population-based cohort study. Eur J Endocrinol. 2017; 176: 1-9.
Page 16/26
35. Xia Y, Hill KE, Li P, XuJ, Zhou D, MotleyAK, et al. Optimization of selenoprotein P and other plasma
selenium biomarkers for the assessment of the selenium nutritional requirement: a placebo-
controlled, double-blind study of selenomethionine supplementation in selenium-decient Chinese
subjects. Am J Clin Nutr. 2010; 92: 525-531.
36. Lutz Schomburg,Marju Orho-Melander,Joachim Struck,Andrea Bergmann,Olle Melander et al.
Selenoprotein-P Deciency Predicts Cardiovascular Disease and Death. Nutrients. 2019; 11: 1852.
37. Cheng P, Wang L, Ning S, LiuZ, LinH, ChenS, et al. Vitamin E intake and risk of stroke: a meta-
analysis. Br J Nutr. 2018; 120: 1181-1188.
38. Beytut E, Yilmaz S, Aksakal M, PolatS, et al. The possible protective effects of vitamin E and
selenium administration in oxidative stress caused by high doses of glucocorticoid administration in
the brain of rats. J Trace Elem Med Biol. 2018; 45: 131-135.
39. Hosnedlova B, Kepinska M, Skalickova S, FernandezC, Ruttkay-NedeckyB, Malevu TD, et al. A
Summary of New Findings on the Biological Effects of Selenium in Selected Animal Species-A
Critical Review. Int J Mol Sci. 2017; 18: 2209.
40. Amani H, Habibey R, Shokri F, HajmiresmailSJ, AkhavanO, MashaghiA, et al. Selenium
nanoparticles for targeted stroke therapy through modulation of inammatory and metabolic
signaling. Sci Rep. 2019; 9: 6044.
41. Mehta SL, Kumari S, Mendelev N, LiPA. Selenium preserves mitochondrial function, stimulates
mitochondrial biogenesis, and reduces infarct volume after focal cerebral ischemia. BMC Neurosci.
2012; 13: 79.
42. Nils Bomer, Niels Grote Beverborg, Martijn FHoes, Koen WStreng, Mathilde Vermeer, Martin M Dokter,
et al. Selenium and outcome in heart failure. Eur J Heart Fail. 2019;6.
43. Alim I, Cauleld JT, Chen Y, SwarupV, GeschwindDH, IvanovaE, et al. Selenium Drives a
Transcriptional Adaptive Program to Block Ferroptosis and Treat Stroke. Cell. 2019; 177: 1262-1279.
Tables
Table 1 Baseline Characteristics of male and female participants.a
Page 17/26
Characteristics Total Male Female
P
value
N n=2510 n=1242 n=1268
Age, y 70.75 ± 8.06 71.42 ± 8.09 70.10 ± 7.98 <0.001
BMI, kg/m226.19 ± 4.10 25.34 ± 3.57 27.02 ± 4.41 <0.001
Current smoking, n (%) 551 (21.95) 547 (44.04) 4 (0.32) <0.001
Current alcohol drinking, n (%) 612 (24.38) 599 (48.23) 13 (1.03) <0.001
Baseline SBP, mmHg 153.25 ± 23.09 150.80 ± 22.34 155.66 ± 23.56 <0.001
Baseline DBP, mmHg 85.25 ± 12.30 86.20 ± 12.50 84.33 ± 12.02 <0.001
Self-reported hypertension, n (%) 1312 (52.27) 553 (44.52) 759 (59.86) <0.001
Self-reported diabetes, n (%) 424 (16.89) 161 (12.96) 263 (20.74) <0.001
Self-reported hyperlipidemia, n (%) 264 (10.52) 108 (8.70) 156 (12.30) 0.003
Self-reported atrial brillation, n (%) 46 (1.83) 26 (2.09) 20 (1.58) 0.335
Hypertension, n (%)b2016 (80.32) 940 (75.68) 1076 (84.86) <0.001
Labor intensity, n (%) <0.001
Mild 1885 (75.10) 887 (71.42) 998 (78.71)
Moderate 488 (19.44) 280 (22.54) 208 (16.40)
Severe 137 (5.46) 75 (6.04) 62 (4.89)
Medication use, n (%)
Antiplatelet drugs 83 (3.31) 47 (3.78) 36 (2.84) 0.186
Lipid-lowering drugs 44 (1.75) 15 (1.21) 29 (2.29) 0.039
Glucose-lowering drugs 301 (11.99) 112 (9.02) 189 (14.91) <0.001
Antihypertensive drugs 1152 (45.90) 476 (38.33) 676 (53.31) <0.001
Laboratory results
TC, mmol/L 5.85 ± 1.20 5.64 ± 1.08 6.05 ± 1.29 <0.001
TG, mmol/L 1.40 ± 0.85 1.21 ± 0.74 1.59 ± 0.91 <0.001
HDL-C, mmol/L 1.63 ± 0.39 1.63 ± 0.42 1.62 ± 0.36 0.558
Glucose, mmol/L 6.26 ± 2.32 6.09 ± 2.09 6.43 ± 2.51 <0.001
tHcy, μmol/L 13.87 ± 7.15 15.26 ± 8.58 12.50 ± 5.03 <0.001
eGFR, mL/min/1.73 m292.49 ± 14.41 92.26 ± 15.06 92.71 ± 13.75 0.428
Page 18/26
Vitamin E, μg/mL 14.04 ± 3.98 12.90 ± 3.27 15.16 ± 4.28 <0.001
Selenium, ng/mL 87.24 ± 18.57 87.68 ± 18.58 86.80 ± 18.55 0.230
a Variables are presented as mean ± SD or n (%). b Hypertension was dened as self-reported history of
hypertension, or use of antihypertensive drugs, or SBP ≥140 mm Hg, or DBP ≥90 mm Hg. Abbreviations:
BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; TC, total cholesterol;
TG, triglycerides; HDL-C, high-density lipoprotein-cholesterol; tHcy, total homocysteine; and eGFR,
estimated glomerular ltration rate.
Table 2 Baseline Characteristics of cases and control participants.a
Page 19/26
Characteristics First stroke cases Non-stroke controls
P
value
N n=1255 n=1255
Age, y 70.75 ± 8.07 70.76 ± 8.06 0.987
Male, n (%) 621 (49.48) 621 (49.48) 1.000
BMI, kg/m226.51 ± 4.42 25.87 ± 3.73 <0.001
Current smoking, n (%) 296 (23.59) 255 (20.32) 0.048
Current alcohol drinking, n (%) 295 (23.51) 317 (25.26) 0.306
Baseline SBP, mmHg 157.17 ± 23.81 149.34 ± 21.66 <0.001
Baseline DBP, mmHg 87.20 ± 12.82 83.31 ± 11.43 <0.001
Self-reported hypertension, n (%) 746 (59.44) 566 (45.10) <0.001
Self-reported diabetes, n (%) 261 (20.80) 163 (12.99) <0.001
Self-reported hyperlipidemia, n (%) 129 (10.28) 135 (10.76) 0.696
Self-reported atrial brillation, n (%) 29 (2.31) 17 (1.35) 0.074
Hypertension, n (%)b1075 (85.66) 941 (74.98) <0.001
Labor intensity, n (%) 0.003
Mild 978 (77.93) 907 (72.27)
Moderate 221 (17.61) 267 (21.27)
Severe 56 (4.46) 81 (6.45)
Medication use, n (%)
Antiplatelet drugs 60 (4.78) 23 (1.83) <0.001
Lipid-lowering drugs 22 (1.75) 22 (1.75) 1.000
Glucose-lowering drugs 193 (15.38) 108 (8.61) <0.001
Antihypertensive drugs 664 (52.91) 488 (38.88) <0.001
Laboratory results
TC, mmol/L 5.84 ± 1.21 5.85 ± 1.20 0.768
TG, mmol/L 1.49 ± 0.92 1.31 ± 0.77 <0.001
HDL-C, mmol/L 1.59 ± 0.38 1.66 ± 0.40 <0.001
Glucose, mmol/L 6.54 ± 2.53 5.98 ± 2.05 <0.001
tHcy, μmol/L 14.26 ± 7.95 13.47 ± 6.22 0.006
Page 20/26
eGFR, mL/min/1.73 m291.69 ± 15.31 93.29 ± 13.40 0.005
Vitamin E, μg/mL 14.13 ± 4.00 13.95 ± 3.95 0.269
Selenium, ng/mL 86.63 ± 17.63 87.84 ± 19.45 0.101
a Variables are presented as mean ± SD or n (%). b Hypertension was dened as self-reported history of
hypertension, or use of antihypertensive drugs, or SBP ≥140 mm Hg, or DBP ≥90 mm Hg. Abbreviations:
BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; TC, total cholesterol;
TG, triglycerides; HDL-C, high-density lipoprotein-cholesterol; tHcy, total homocysteine; and eGFR,
estimated glomerular ltration rate.
Table 3 Risk of rst stroke (total and subtypes) associated with plasma selenium concentrations in total
participants.a
Page 21/26
Selenium, ng/mL Cases/controls Model 1 Model 2
OR (95% CI)
P
value OR (95% CI)
P
value
First total stroke
Tertiles
T1 (<79.1) 420/417
Ref. Ref.
T2 (79.1 to <94.1) 434/402 1.07 (0.88, 1.30) 0.516 1.02 (0.82, 1.26) 0.891
T3 (≥94.1) 401/436 0.90 (0.73, 1.11) 0.317 0.77 (0.61, 0.97) 0.027
Categories
T1-T2 (<94.1) 854/819 Ref. Ref.
T3 (≥94.1) 401/436 0.87 (0.73, 1.04) 0.119 0.76 (0.63, 0.93) 0.007
Ischemic stroke
Tertiles
T1 (<79.1) 360/363
Ref. Ref.
T2 (79.1 to <94.1) 372/340 1.10 (0.89, 1.35) 0.376 1.05 (0.83, 1.33) 0.672
T3 (≥94.1) 347/376 0.92 (0.74, 1.14) 0.444 0.78 (0.60, 0.99) 0.045
Categories
T1-T2 (<94.1) 732/703
Ref. Ref.
T3 (≥94.1) 347/376 0.87 (0.72, 1.06) 0.163 0.75 (0.61, 0.93) 0.009
Hemorrhagic stroke
Tertiles
T1 (<79.1) 58/54
Ref. Ref.
T2 (79.1 to <94.1) 60/61 0.88 (0.50, 1.57) 0.673 0.76 (0.38, 1.55) 0.455
T3 (≥94.1) 53/56 0.85 (0.47, 1.54) 0.587 0.72 (0.35, 1.48) 0.371
Categories
T1-T2 (<94.1) 118/115
Ref. Ref.
T3 (≥94.1) 53/56 0.92 (0.57, 1.47) 0.718 0.85 (0.48, 1.50) 0.577
a ORs of rst stroke (total), rst ischemic and hemorrhagic stroke in relation to plasma selenium (tertiles)
were analyzed using conditional logistic regression models. Model 1 is conditioned on the matching
factors of age and sex; Model 2 is conditioned on the matching factors of age and sex, as well as
Page 22/26
adjusted for BMI, baseline SBP, baseline DBP, smoking status, alcohol consumption, labor intensity,
baseline total homocysteine, vitamin E, fasting glucose, estimated glomerular ltration rate (eGFR),
antiplatelet drugs, lipoprotein-lowering drugs, glucose-lowering drugs, antihypertensive drugs, self-
reported hypertension, self-reported atrial brillation, self-reported diabetes, and self-reported
hyperlipidemia.
Table 4. The association of plasma selenium with risk of rst stroke (total and subtypes) by sex.a
Page 23/26
bMale participants Female participants
Cases/controls Model 1 Model 2 Cases/controls Model 1 Model 2
OR (95% CI)
P
value OR (95% CI)
P
value OR
(95% CI)
P
value
OR
(95% CI)
P
value
First total stroke
Selenium Tertiles
T1 (<79.1
ng/mL) 206/193
Ref. Ref.
214/224
Ref. Ref.
T2 (79.1
to <94.1) 218/183 1.12 (0.84,
1.49) 0.452 1.11 (0.80,
1.54) 0.519 216/219 1.04
(0.80,
1.36)
0.764
0.88
(0.65,
1.19)
0.408
T3
(≥94.1) 197/245 0.73 (0.55,
0.98) 0.034 0.67 (0.48,
0.93) 0.017 204/191 1.14
(0.85,
1.54)
0.380
0.85
(0.61,
1.19)
0.353
Selenium Categories
T1-T2
(<94.1) 424/376
Ref. Ref.
430/443
Ref. Ref.
T3
(≥94.1) 197/245 0.69 (0.54,
0.89) 0.003 0.63 (0.48,
0.83) 0.001 204/191 1.12
(0.86,
1.44)
0.399
0.92
(0.69,
1.22)
0.563
First ischemic stroke
Selenium Tertiles
T1 (<79.1
ng/mL) 179/171
Ref. Ref.
181/192
Ref. Ref.
T2 (79.1
to <94.1) 191/157 1.16 (0.85,
1.57) 0.343 1.17 (0.82,
1.66) 0.387 181/183 1.05
(0.79,
1.40)
0.716
0.90
(0.64,
1.25)
0.514
T3
(≥94.1) 168/210 0.74 (0.55,
1.01) 0.058 0.67 (0.47,
0.95) 0.027 179/166 1.17
(0.85,
1.60)
0.330
0.84
(0.58,
1.21)
0.347
Selenium Categories
T1-T2
(<94.1) 370/328
Ref. Ref.
362/375
Ref. Ref.
T3 168/210 0.69 (0.53, 0.61 (0.45, 179/166 1.14 0.89
Page 24/26
(≥94.1) 0.90) 0.006 0.83) 0.001 (0.86,
1.50)
0.362
(0.65,
1.22)
0.479
First hemorrhagic stroke
Selenium Tertiles
T1 (<79.1
ng/mL) 26/22
Ref. Ref.
32/32
Ref. Ref.
T2 (79.1
to <94.1) 26/26 0.81 (0.33,
2.00) 0.646 0.88 (0.25,
3.04) 0.840 34/35 0.98
(0.46,
2.12)
0.962
0.48
(0.17,
1.37)
0.172
T3
(≥94.1) 29/33 0.72 (0.32,
1.60) 0.420 0.80 (0.25,
2.58) 0.714 24/23 1.05
(0.43,
2.57)
0.919
0.58
(0.17,
1.92)
0.368
Selenium Categories
T1-T2
(<94.1) 52/48
Ref. Ref.
66/67
Ref. Ref.
T3
(≥94.1) 29/33 0.80 (0.41,
1.54) 0.506 0.86 (0.33,
2.26) 0.761 24/23 1.06
(0.54,
2.10)
0.862
0.95
(0.36,
2.46)
0.910
aModel 1 is conditioned on the matching factors of age and sex; Model 2 is conditioned on the matching
factors of age and sex, as well as adjusted for BMI, baseline SBP, baseline DBP, smoking status, alcohol
consumption, labor intensity, baseline total homocysteine, vitamin E, fasting glucose, estimated
glomerular ltration rate (eGFR), antiplatelet drugs, lipoprotein-lowering drugs, glucose-lowering drugs,
antihypertensive drugs, self-reported hypertension, self-reported diabetes, self-reported atrial brillation,
and self-reported hyperlipidemia. b Adjusted
P
-interaction between sex and plasma selenium (T3, ≥94.1
ng/mL
vs.
T1-2, <94.1 ng/mL) on rst total stroke=0.029. Abbreviations: T, tertile; OR,odds ratio; CI,
condence interval.
Table 5 Stratied analysis of the association between plasma selenium concentrations (T3, ≥94.1 ng/mL
vs.
T1-2, <94.1 ng/mL) and incident risk of rst total stroke in males.
Page 25/26
Subgroups No. of cases / No. of controls aAdjusted
Model
P
for
interaction
Selenium≥94.1
ng/mL Selenium<94.1
ng/mL OR (95% CI)
Age, y 0.288
<70 98/110 173/161 0.73 (0.50,
1.08)
≥70 99/135 251/215 0.54 (0.38,
0.76)
Body mass index,
kg/m2
0.067
<24 44/89 151/152 0.39 (0.24,
0.63)
≥24 153/156 273/224 0.75 (0.55,
1.02)
Current smoking 0.739
No 95/136 233/231 0.72 (0.51,
1.01)
Yes 102/109 191/145 0.54 (0.36,
0.80)
Current alcohol
drinking 0.230
No 98/106 232/207 0.74 (0.51,
1.06)
Yes 99/139 192/169 0.53 (0.37,
0.77)
SBP, mmHg 0.759
<140 46/92 124/156 0.60 (0.38,
0.96)
≥140 151/153 300/220 0.64 (0.47,
0.88)
Glucose, mmol/L 0.898
<6.1 108/168 274/284 0.66 (0.48,
0.90)
≥6.1 or diabetesb89/77 150/92 0.66 (0.42,
1.02)
TC, mmol/L 0.402
Page 26/26
<5.78 95/107 262/236 0.77 (0.54,
1.10)
≥5.78 102/138 162/140 0.53 (0.36,
0.78)
TG, mmol/L 0.611
<1.17 115/160 246/255 0.71 (0.52,
0.98)
≥1.17 82/85 178/121 0.55 (0.36,
0.84)
eGFR,
mL/min/1.73m2
0.753
<90 64/68 173/141 0.71 (0.45,
1.10)
≥90 133/177 251/235 0.61 (0.45,
0.83)
tHcy, μmol/L 0.117
<12.5 89/105 129/142 0.81 (0.54,
1.22)
≥12.5 108/140 293/234 0.55 (0.39,
0.76)
Vitamin E, μg/mL 0.007
<13.5 117/132 270/277 0.85 (0.62,
1.17)
≥13.5 80/113 154/99 0.39 (0.25,
0.60)
a ORs of rst total stroke in relation to serum selenium levels were calculated using multivariate logistic
regression models. Each subgroup analysis adjusted, if not stratied,for age, BMI, baseline SBP, baseline
DBP, smoking status, alcohol consumption, labor intensity, baseline total homocysteine, vitamin E, fasting
glucose, estimated glomerular ltration rate (eGFR), antiplatelet drugs, lipoprotein-lowering drugs,
glucose-lowering drugs, antihypertensive drugs, self-reported hypertension, self-reported diabetes, self-
reported atrial brillation, and self-reported hyperlipidemia. b Diabetes was dened as self-reported history
of diabetes mellitus, or use of anti-diabetic medications, or fasting glucose ≥7.0 mmol/L. Abbreviations:
TC, total cholesterol; T, tertile; OR, odds ratio; CI, condence interval.