Relationship between Serum Selenium Levels and Hypertension: Findings from the National Health and Nutrition Examination Survey, 2009–2018

Abstract

Objective: This study was aimed at evaluating the association between serum selenium (Se) levels and hypertension in United States (US) adults.

Methods: Data were extracted from the United States National Health and Nutrition Examination Survey (NHANES, 2009–2018). Multivariate logistic regression and subgroup analysis were conducted to examine associations between Se and hypertension.

Results: Patients with hypertension had higher serum Se levels than healthy participants (P = 0.01017). After adjustment for demographic and health behaviors, serum Se levels were significantly associated with systolic blood pressure in women (OR = 0.014; 95% CI: 0.001–0.027; P = 0.04174). Serum Se levels were significantly associated with diastolic blood pressure (DBP) in both women and men. In subgroup analysis, serum Se levels were significantly associated with higher diastolic blood pressure (P for trend <0.001).

Conclusion: In summary, this analysis supports a nonlinear association between serum Se levels and the prevalence of hypertension, and higher serum Se levels were associated with higher DBP.

Main article text

Introduction

Hypertension, also known as high blood pressure, poses a major public health burden worldwide, because of the associated high risk of cardiovascular disease and mortality. In recent decades, the global prevalence of hypertension has steadily increased, and the number of individuals affected by hypertension is expected to reach 1.5 billion by 2025, according to a brief report released by the World Health Organization. Nearly one in three adults >18 years of age in the United States (US) has hypertension [1–4]. Multiple studies have demonstrated that hypertension is a multifactorial disease; therefore, managing hypertension risk factors is critical to decrease disease burden and improve the quality of life and health prospects of patients with hypertension. Many widely recognized risk factors, including obesity, cigarette smoking, unhealthful diets, and physical inactivity, have attracted substantial attention. However, the roles of several trace elements in the pathogenesis of hypertension have not been well researched [1, 4–6].

Serum selenium (Se), an essential trace element in the human body [7], plays important roles in the antioxidant response via glutathione peroxidases and selenoenzymes [6]. Oxidative stress is a primary contributor to numerous diseases including cancer, non-alcoholic fatty liver disease, diabetes, hypertension, and cardiovascular disease [7–13]. Se has been suggested to be a protective factor associated with blood pressure control and hypertension prevention [14]. The differences in results among studies suggest that patient characteristics may influence the response to Se. However, findings from studies evaluating the potential association between serum Se levels and hypertension have been inconsistent: whereas some studies have reported an inverse association, others have reported no association or a positive association. Epidemiological studies by Laclaustra et al. found a correlation between low blood Se levels and the development of hypertension in humans [15]. Kuruppu et al. have found no conclusive evidence supporting an association between Se and hypertension in a recent systematic review [6, 15]. Some observational studies have reported that high Se levels may be associated with elevated cardiovascular disease risk, particularly in the older population [16–18]. Comprehensive study of the relationship between Se and hypertension would require randomized controlled trials and prospective studies with sufficiently large sample sizes to be conducted in populations with different Se levels. Nawrot et al. have reported that Se deficiency might be an underestimated risk factor in the development of hypertension in European men [19]. A study using data from the National Health and Nutrition Examination Survey (NHANES) has indicated that high values of serum Se, including high-normal values, may lead to hypertension, regardless of age [18]. This study was aimed at assessing the relationship between serum Se levels and hypertension through a cross-sectional analysis, to aid in management of risk factors for the development of hypertension.

Methods

NHANES Dataset

The data analyzed in this study were obtained from NHANES, a large continuing survey examining public health and nutrition in the civilian, non-institutionalized, general population in the US. Data from five 2-year cycles of NHANES surveys (2009–2010, 2011–2012, 2013–2014, 2015–2016, and 2017–2018) were collected and combined to create a dataset from 2009 to 2018. The survey protocols were approved by the National Center for Health Statistics, and all participants provided written consent to use their anonymized data and information for further research [20].

Study Population

A total of 49,693 participants from NHANES 2009–2018 were enrolled to determine research eligibility, and 9641 participants <18 years of age were excluded. Individuals without the requisite Se data (n = 24,058) and missing covariate data (n = 296) were excluded. Thus, 15,698 participants were included in the final analysis.

The study examiners measured blood pressure according to their protocol. Participants were defined as hypertensive if they met one or more of the following criteria: systolic blood pressure (SBP) ≥130 mmHg and/or diastolic blood pressure (DBP) ≥80 mmHg [21]; self-reported, physician-diagnosed hypertension; self-reported high blood pressure diagnosed by a physician; and/or self-reported treatment with antihypertensive drugs.

Measurement of Serum Se Levels

The protocol for measuring serum Se levels has been described in detail elsewhere. Briefly, blood samples were collected at baseline. After centrifugation, serum aliquots were transported at 4°C to the Trace Elements Laboratory at the Wadsworth Center of the New York State Department of Health (New York, USA) and measured with inductively coupled plasma-dynamic reaction cell-mass spectrometry, or stored at −20°C or −70°C until analysis [11, 18].

Covariates

Covariates selected for this study were based on clinical experience, and all data were extracted from the NHANES database. The covariates included demographic factors (age, sex, and racial/ethnic group); social-behavioral factors (smoking status, educational status, and alcohol consumption); clinical factors (blood pressure); and laboratory findings. Details for all covariate processes are available from the NHANES website.

Statistical Analysis

R software version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria) and EmpowerStats software version 2.2 were used for all analyses. Participant variables were grouped according to the presence or absence of hypertension. Continuous variables are expressed as mean and standard deviation (SD), whereas categorical variables are expressed as frequencies or percentages (weighted). Student’s t-test, one-way analysis of variance, and chi-squared tests were used to compare and determine statistical differences among subgroups. Multivariate logistic regression models were used to explore the association between serum Se level and blood pressure. No covariates were adjusted for in model 1; model 2 was adjusted for sex and age; and model 3 was adjusted for sex, age, race/ethnicity, marital status, educational level, total number of individuals in the household, smoking, alcohol use, aspartate aminotransferase (AST), alanine aminotransferase (ALT), triglycerides (TG), total cholesterol (TC), serum creatinine, blood urea nitrogen, and plasma glucose level. Differences with P < 0.05 were considered statistically significant.

Results

Baseline Characteristics

The survey-weighted characteristics of the participants are summarized in Table 1. A total of 15,698 participants with complete data for serum Se, blood pressure, and other covariates were divided into two groups according to blood pressure: 8559 had hypertension, and 7139 had normal blood pressure. The mean (± SD) age of all participants in this study was 48.57 ± 18.47 years; 7661 were men, and 8037 were women. The mean serum Se level was 193.291 μg/L, and patients with hypertension exhibited higher serum Se levels than those with normal blood pressure (193.998 ± 28.654 vs. 192.867 ± 25.762 μg/L; P = 0.01017). Compared with healthy participants, participants with hypertension had higher ALT, AST, ALP, uric acid, TC, TG, serum creatinine, blood urea nitrogen, and plasma glucose levels. The participants included in this study were divided into four quartiles according to serum Se level. The characteristics of all participants in each quartile are summarized in Table 2. Participants included in the hypertension population were divided into four quartiles according to serum Se levels. The characteristics of the participants with hypertension in each of these quartiles are summarized in Tables 3 and 4. Many covariates significantly differed among these four quartiles, including sex, age, race, education level, smoking status, alcohol use, blood pressure, ALT, AST, APL, uric acid, TC, TG, serum creatinine, and plasma glucose. The weighted characteristics of the study population by serum Se quartile among hypertensive and normotensive participants are presented in Tables 3 and 4.

Table 1

Comparison between Hypertension and Non-Hypertension in the Overall Population.

Characteristics	Overall (n = 15,698)	Non-Hypertension (n = 7139)	Hypertension (n = 8559)	P value*
Sex				<0.001 †
Male	7661	3206 (44.908%)	4455 (52.050%)
Female	8037	3933 (55.092%)	4104 (47.950%)
Age	48.57 ± 18.47	38.68 ± 16.06	56.53 ± 16.18	<0.001 †
Race/ethnicity				<0.001 †
Mexican American	2092	1115 (15.62%)	977 (11.41%)
Other Hispanic	1624	805 (11.28%)	819 (9.57%)
Non-Hispanic White	5758	2564 (35.92%)	3194 (37.32%)
Non-Hispanic Black	3600	1291 (18.08%)	2309 (26.98%)
Other race	2624	1364 (19.11%)	1260 (14.72%)
Education level				<0.001 †
Below high school	3236	1232 (17.26%)	2004 (23.41%)
High school	3361	1334 (18.69%)	2027 (23.68%)
College or above	8328	3908 (54.74%)	4420 (51.64%)
Other	773	665 (9.32%)	108 (1.26%)
Total number of people in the household				<0.001 †
1	2134	615 (8.6%)	1519 (17.7%)
2	4652	1696 (23.8%)	2956 (34.5%)
3	2849	1462 (20.5%)	1387 (16.2%)
4	2579	1430 (20.0%)	1149 (13.4%)
5	1758	979 (13.7%)	779 (9.1%)
6	912	501 (7.0%)	411 (4.8%)
7 or more people in the household	814	456 (6.4%)	358 (4.2%)
Smoking status				<0.001 †
Never smoke	2878	1293 (18.11%)	1585 (18.52%)
Always smoke	3562	1170 (16.39%)	2392 (27.95%)
Still smoke	9258	4676(65.50%)	4582 (53.53%)
Alcohol use				<0.001 †
Never drink	998	282 (3.95%)	716 (8.37%)
Always drink	3285	1505 (21.08%)	1780 (20.80%)
Still drink	11415	5352 (74.97%)	6063 (70.84%)
Marital status				<0.001 †
Living with partner	8742	3840 (53.79%)	4902 (57.27%)
Separated	6189	2632 (36.87%)	3557 (41.56%)
Other	767	667 (9.34%)	100 (1.17%)
Systolic blood pressure (mmHg)	124.41 ± 18.666	111.982 ± 9.171	134.830 ± 18.401	<0.001 †
Diastolic blood pressure (mmHg)	70.789 ± 12.204	66.583 ± 8.621	74.806 ± 13.475	<0.001 †
Serum selenium (μg/L)	193.291 ± 26.992	192.867 ± 25.762	193.998 ± 28.654	0.01017*
ALT (U/L)	23.96 ± 20.787	22.035 ± 16.170	24.671 ± 24.971	<0.001 †
AST (U/L)	24.441 ± 17.23	23.196 ± 14.683	25.173 ± 18.382	<0.001 †
APL (U/L)	71.976 ± 25.327	66.827 ± 22.209	74.210 ± 26.599	<0.001 †
Uric acid (μmol/L)	323.41 ± 85.317	303.429 ± 77.529	343.566 ± 87.834	<0.001 †
TG (mmol/L)	1.671 ± 1.339	1.428 ± 1.128	1.781 ± 1.374	<0.001 †
TC (mmol/L)	4.893 ± 1.068	4.793 ± 1.018	4.957 ± 1.101	<0.001 †
Serum creatinine (μmol/L)	79.667 ± 39.820	73.609 ± 20.964	86.442 ± 52.264	<0.001 †
Blood urea nitrogen (mmol/L)	4.987 ± 2.148	4.442 ± 1.550	5.370 ± 2.475	<0.001 †
Plasma glucose (mmol/L)	5.711 ± 2.147	5.261 ± 1.578	6.019 ± 2.382	<0.001 †

Aspartate Aminotransferase: AST; Alanine Aminotransferase: ALT; Alkaline Phosphatase: ALP; Triglycerides: TG; Total Cholesterol: TC.

*P < 0.05; ^†P < 0.01.

Table 2

Weighted Characteristics of the Study Population based on Serum Selenium Quartiles in the Overall Population.

Selenium quartiles (μg/L)	Q1	Q2	Q3	Q4	P value
Sex					<0.001 †
Male	1670 (42.559%)	1852 (47.185%)	2000 (50.968%)	2139 (54.497%)
Female	2254 (57.441%)	2073 (52.815%)	1924 (49.032%)	1786 (45.503%)
Age	49.674 ± 19.305	48.158 ± 18.560	47.970 ± 18.142	48.517 ± 17.830	<0.001 †
Race/ethnicity					<0.001 †
Mexican American	465 (11.850%)	541 (13.783%)	549 (13.991%)	537 (13.682%)
Other Hispanic	457 (11.646%)	436 (11.108%)	400 (10.194%)	331 (8.433%)
Non-Hispanic White	1357 (34.582%)	1367 (34.828%)	1503 (38.303%)	1531 (39.006%)
Non-Hispanic Black	1105 (28.160%)	938 (23.898%)	829 (21.126%)	728 (18.548%)
Other race	540 (13.761%)	643 (16.382%)	643 (16.386%)	798 (20.331%)
Education level					<0.001 †
Below high school	947 (24.134%)	804 (20.484%)	739 (18.833%)	746 (19.006%)
High school	884 (22.528%)	830 (21.146%)	834 (21.254%)	813 (20.713%)
College or above	1894 (48.267%)	2091 (53.274%)	2162 (55.097%)	2181 (55.567%)
Other	199 (5.071%)	200 (5.096%)	189 (4.817%)	185 (4.713%)
Total number of people in the household					0.365
1	592 (15.087%)	522 (13.299%)	489 (12.462%)	531 (13.529%)
2	1136 (28.950%)	1137 (28.968%)	1208 (30.785%)	1171 (29.834%)
3	687 (17.508%)	752 (19.159%)	705 (17.966%)	705 (17.962%)
4	637 (16.233%)	657 (16.739%)	643 (16.386%)	642 (16.357%)
5	435 (11.086%)	434 (11.057%)	441 (11.239%)	448 (11.414%)
6	223 (5.683%)	225 (5.732%)	242 (6.167%)	222 (5.656%)
7 or more people in the household	214 (5.454%)	198 (5.045%)	196 (4.995%)	206 (5.248%)
Smoking status					<0.001 †
Never smoke	833 (21.228%)	728 (18.548%)	672 (17.125%)	645 (16.433%)
Always smoke	841 (21.432%)	841 (21.427%)	921 (23.471%)	959 (24.433%)
Still smoke	2250 (57.339%)	2356 (60.025%)	2331 (59.404%)	2321 (59.134%)
Alcohol use					<0.001 †
Never drink	312 (7.951%)	241 (6.140%)	234 (5.963%)	211 (5.376%)
Always drink	915 (23.318%)	847 (21.580%)	792 (20.183%)	731 (18.624%)
Still drink	2697 (68.731%)	2837 (72.280%)	2898 (73.853%)	2983 (76.000%)
Marital status					<0.001 †
Living with partner	2024 (51.580%)	2128 (54.217%)	2284 (58.206%)	2306 (58.752%)
Separated	1702 (43.374%)	1596 (40.662%)	1457 (37.130%)	1434 (36.535%)
Other	198 (5.046%)	201 (5.121%)	183 (4.664%)	185 (4.713%)
Systolic blood pressure (mmHg)	124.626 ± 20.081	124.102 ± 18.835	124.042 ± 18.210	124.873 ± 17.431	0.0001 †
Diastolic blood pressure (mmHg)	69.285 ± 12.300	70.554 ± 12.113	71.337 ± 12.146	71.965 ± 12.097	<0.001 †
ALT (U/L)	21.478 ± 15.532	23.565 ± 26.328	23.982 ± 17.863	26.815 ± 21.462	<0.001 †
AST (U/L)	24.160 ± 19.545	24.060 ± 15.469	23.989 ± 17.890	25.556 ± 15.643	<0.001 †
APL (U/L)	74.460 ± 30.412	72.375 ± 23.684	70.798 ± 23.372	70.271 ± 22.885	<0.001 †
Uric acid (μmol/L)	314.593 ± 87.959	322.423 ± 85.628	324.801 ± 82.431	331.819 ± 84.301	<0.001 †
TG (mmol/L)	1.466 ± 1.067	1.606 ± 1.290	1.713 ± 1.396	1.899 ± 1.525	<0.001 †
TC (mmol/L)	4.684 ± 1.020	4.859 ± 1.040	4.926 ± 1.034	5.102 ± 1.131	<0.001 †
Serum creatinine (mmol/L)	82.773 ± 56.456	79.189 ± 40.002	78.728 ± 30.840	77.970 ± 24.328	0.398
Blood urea nitrogen (mmol/L)	5.121 ± 2.624	4.988 ± 2.092	4.950 ± 1.976	4.888 ± 1.806	0.023*
Plasma glucose (mmol/L)	5.613 ± 1.952	5.648 ± 2.043	5.723 ± 2.240	5.861 ± 2.323	<0.001 †

Aspartate Aminotransferase: AST; Alanine Aminotransferase: ALT; Alkaline Phosphatase: ALP; Triglycerides: TG; Total Cholesterol: TC.

*P < 0.05; ^†P < 0.01.

Table 3

Weighted Characteristics of the Study Population based on Serum Selenium Quartiles in the Population with Hypertension.

Selenium quartiles (μg/L)	Q1	Q2	Q3	Q4	P value
Sex					<0.001 †
Male	1000 (46.970%)	1065 (51.006%)	1116 (53.042%)	1274 (56.926%)
Female	1129 (53.030%)	1023 (48.994%)	988 (46.958%)	964 (43.074%)
Age	59.100 ± 16.198	56.769 ± 16.154	56.013 ± 15.937	55.157 ± 16.108	<0.001 †
Race/ethnicity					<0.001 †
Mexican American	207 (9.723%)	244 (11.686%)	248 (11.787%)	278 (12.422%)
Other Hispanic	204 (9.582%)	229 (10.967%)	206 (9.791%)	180 (8.043%)
Non-Hispanic White	779 (36.590%)	730 (34.962%)	813 (38.641%)	872 (38.963%)
Non-Hispanic Black	711 (33.396%)	590 (28.257%)	532 (25.285%)	476 (21.269%)
Other race	228 (10.709%)	295 (14.128%)	305 (14.496%)	432 (19.303%)
Education level					<0.001 †
Below high school	602 (28.276%)	489 (23.420%)	447 (21.245%)	466 (20.822%)
High school	547 (25.693%)	483 (23.132%)	490 (23.289%)	507 (22.654%)
College or above	959 (45.045%)	1088 (52.107%)	1137 (54.040%)	1236 (55.228%)
Other	21 (0.986%)	28 (1.341%)	30 (1.426%)	29 (1.296%)
Total number of people in the household					0.481
1	420 (19.728%)	369 (17.672%)	354 (16.825%)	376 (16.801%)
2	728 (34.194%)	724 (34.674%)	747 (35.504%)	757 (33.825%)
3	328 (15.406%)	359 (17.193%)	336 (15.970%)	364 (16.265%)
4	269 (12.635%)	286 (13.697%)	273 (12.975%)	321 (14.343%)
5	189 (8.877%)	176 (8.429%)	206 (9.791%)	208 (9.294%)
6	101 (4.744%)	94 (4.502%)	105 (4.990%)	111 (4.960%)
7 or more people in the household	94 (4.415%)	80 (3.831%)	83 (3.945%)	101 (4.513%)
Smoking status					<0.001 †
Never smoke	476 (22.358%)	385 (18.439%)	356 (16.920%)	368 (16.443%)
Always smoke	587 (27.572%)	566 (27.107%)	581 (27.614%)	658 (29.401%)
Still smoke	1066 (50.070%)	1137 (54.454%)	1167 (55.466%)	1212 (54.155%)
Alcohol use					<0.001 †
Never drink	221 (10.380%)	176 (8.429%)	163 (7.747%)	156 (6.971%)
Always drink	465 (21.841%)	447 (21.408%)	421 (20.010%)	447 (19.973%)
Still drink	1443 (67.778%)	1465 (70.163%)	1520 (72.243%)	1635 (73.056%)
Marital status					<0.001 †
Living with partner	1107 (51.996%)	1161 (55.603%)	1263 (60.029%)	1371 (61.260%)
Separated	1001 (47.017%)	900 (43.103%)	817 (38.831%)	839 (37.489%)
Other	21 (0.986%)	27 (1.293%)	24 (1.141%)	28 (1.251%)
Systolic blood pressure (mmHg)	135.803 ± 19.918	135.072 ± 18.315	134.217 ± 18.011	133.820 ± 16.859	0.016*
Diastolic blood pressure (mmHg)	72.651 ± 13.622	74.277 ± 13.446	75.326 ± 13.318	75.508 ± 13.175	<0.001 †
ALT (U/L)	21.836 ± 14.643	25.001 ± 33.480	24.957 ± 19.984	27.876 ± 20.746	<0.001 †
AST (U/L)	24.602 ± 14.881	25.151 ± 18.537	24.923 ± 22.755	26.282 ± 15.947	<0.001 †
APL (U/L)	78.674 ± 32.746	75.977 ± 24.517	73.363 ± 24.809	72.351 ± 24.269	<0.001 †
Uric acid (μmol/L)	335.515 ± 91.284	341.640 ± 87.405	339.731 ± 84.357	346.789 ± 86.452	<0.001 †
TG (mmol/L)	1.575 ± 1.135	1.790 ± 1.454	1.870 ± 1.434	2.091 ± 1.622	<0.001 †
TC (mmol/L)	4.733 ± 1.067	4.940 ± 1.071	5.006 ± 1.069	5.180 ± 1.154	<0.001 †
Serum creatinine (μmol/L)	92.132 ± 71.905	85.062 ± 51.613	83.119 ± 36.343	81.134 ± 28.223	0.001 †
Blood urea nitrogen (mmol/L)	5.679 ± 3.103	5.437 ± 2.403	5.319 ± 2.197	5.194 ± 1.966	0.021*
Plasma glucose (mmol/L)	5.935 ± 2.169	6.044 ± 2.437	6.084 ± 2.562	6.195 ± 2.564	<0.001 †

Aspartate Aminotransferase: AST; Alanine Aminotransferase: ALT; Alkaline Phosphatase: ALP; Triglycerides: TG; Total Cholesterol: TC.

*P < 0.05; ^†P < 0.01.

Table 4

Weighted Characteristics of the Study Population based on Serum Selenium Quartiles of the Non-Hypertension Population.

Selenium quartiles (μg/L)	Q1	Q2	Q3	Q4	P value
Sex					<0.001 †
Male	670 (37.326%)	787 (42.842%)	884 (48.571%)	865 (51.274%)
Female	1125 (62.674%)	1050 (57.158%)	936 (51.429%)	822 (48.726%)
Age	38.494 ± 16.531	38.370 ± 16.107	38.671 ± 15.965	39.709 ± 16.110	0.021*
Race/ethnicity					<0.001 †
Mexican American	258 (14.373%)	297 (16.168%)	301 (16.538%)	259 (15.353%)
Other Hispanic	253 (14.095%)	207 (11.268%)	194 (10.659%)	151 (8.951%)
Non-Hispanic White	578 (32.201%)	637 (34.676%)	690 (37.912%)	659 (39.063%)
Non-Hispanic Black	394 (21.950%)	348 (18.944%)	297 (16.319%)	252 (14.938%)
Other race	312 (17.382%)	348 (18.944%)	338 (18.571%)	366 (21.695%)
Education level					0.251
Below high school	345 (19.220%)	315 (17.148%)	292 (16.044%)	280 (16.598%)
High school	337 (18.774%)	347 (18.889%)	344 (18.901%)	306 (18.139%)
College or above	935 (52.089%)	1003 (54.600%)	1025 (56.319%)	945 (56.017%)
Other	178 (9.916%)	172 (9.363%)	159 (8.736%)	156 (9.247%)
Total number of people in the household					0.634
1	172 (9.582%)	153 (8.329%)	135 (7.418%)	155 (9.188%)
2	408 (22.730%)	413 (22.482%)	461 (25.330%)	414 (24.541%)
3	359 (20.000%)	393 (21.394%)	369 (20.275%)	341 (20.213%)
4	368 (20.501%)	371 (20.196%)	370 (20.330%)	321 (19.028%)
5	246 (13.705%)	258 (14.045%)	235 (12.912%)	240 (14.226%)
6	122 (6.797%)	131 (7.131%)	137 (7.527%)	111 (6.580%)
7 or more people in the household	120 (6.685%)	118 (6.424%)	113 (6.209%)	105 (6.224%)
Smoking status					<0.001 †
Never smoke	357 (19.889%)	343 (18.672%)	316 (17.363%)	277 (16.420%)
Always smoke	254 (14.150%)	275 (14.970%)	340 (18.681%)	301 (17.842%)
Still smoke	1184 (65.961%)	1219 (66.358%)	1164 (63.956%)	1109 (65.738%)
Alcohol use					<0.001 †
Never drink	91 (5.070%)	65 (3.538%)	71 (3.901%)	55 (3.260%)
Always drink	450 (25.070%)	400 (21.775%)	371 (20.385%)	284 (16.835%)
Still drink	1254 (69.861%)	1372 (74.687%)	1378 (75.714%)	1348 (79.905%)
Marital status					0.057
Living with partner	917 (51.086%)	967 (52.640%)	1021 (56.099%)	935 (55.424%)
Separated	701 (39.053%)	696 (37.888%)	640 (35.165%)	595 (35.270%)
Other	177 (9.861%)	174 (9.472%)	159 (8.736%)	157 (9.306%)
Systolic blood pressure (mmHg)	111.370 ± 9.331	111.632 ± 9.192	112.278 ± 9.052	113.003 ± 9.100	<0.001 †
Diastolic blood pressure (mmHg)	65.291 ± 9.019	66.323 ± 8.625	66.725 ± 8.570	67.266 ± 8.457	<0.001 †
ALT (U/L)	21.054 ± 16.520	21.932 ± 14.222	22.855 ± 14.970	25.407 ± 22.305	<0.001 †
AST (U/L)	23.635 ± 23.920	22.820 ± 10.860	22.910 ± 9.459	24.594 ± 15.182	0.003 †
APL (U/L)	69.464 ± 26.542	68.281 ± 22.002	67.832 ± 21.215	67.511 ± 20.594	0.402
Uric acid (μmol/L)	289.778 ± 76.802	300.580 ± 78.033	307.542 ± 76.625	311.961 ± 77.015	<0.001 †
TG (mmol/L)	1.335 ± 0.964	1.397 ± 1.035	1.532 ± 1.327	1.645 ± 1.346	<0.001 †
TC (mmol/L)	4.624 ± 0.957	4.766 ± 0.995	4.834 ± 0.985	5.000 ± 1.092	<0.001 †
Serum creatinine (μmol/L)	71.672 ± 24.688	72.515 ± 17.550	73.652 ± 21.821	73.772 ± 17.020	<0.001 †
Blood urea nitrogen (mmol/L)	4.460 ± 1.683	4.478 ± 1.516	4.524 ± 1.582	4.482 ± 1.473	0.661
Plasma glucose (mmol/L)	5.231 ± 1.577	5.198 ± 1.338	5.306 ± 1.704	5.418 ± 1.871	<0.001 †

Aspartate Aminotransferase: AST; Alanine Aminotransferase: ALT; Alkaline Phosphatase: ALP; Triglycerides: TG; Total Cholesterol: TC.

*P < 0.05; ^†P < 0.01.

Association between Blood Pressure and Serum Se Level

Multivariate logistic regression models were used to determine the relationships of serum Se level with SBP and DBP. Model 1 was a non-adjusted model; model 2 was adjusted for sex and age; and model 3 was adjusted for the parameters in model 2 and other parameters including race/ethnicity, marital status, total number of individuals in the household, smoking status, alcohol use, TG, TC, serum creatinine, uric acid, and plasma glucose level.

Among all participants, serum Se levels were positively correlated with SBP in model 2 (OR 0.013); however, this association was not significant in models 1 and 2 (OR 0.007 and 0.005, respectively). A negative correlation was observed between serum Se and DBP in models 1, 2, and 3 (OR 0.036, 0.034, and 0.024, respectively).

In the subgroup analysis, serum Se levels were significantly associated with SBP in model 1 and model 2 in women. Serum Se levels were significantly associated with DBP in both women and men. In Mexican Americans, serum Se was significantly associated with SBP in model 1, whereas higher serum Se levels were significantly associated with higher DBP. Participant data are summarized in Tables 5 and 6.

Table 5

Association between Serum Selenium (μg/L) and Systolic Blood Pressure (mmHg).

	Model 1 β/OR (95%CI) P value	Model 2 β/OR (95%CI) P value	Model 3 β/OR (95%CI) P value
Serum selenium (μg/L)	0.007 (−0.004, 0.018) 0.19586	0.013 (0.003, 0.022) 0.00886 †	0.005 (−0.005, 0.014) 0.31985
Subgroup analysis, stratified by sex
Male	−0.011 (−0.026, 0.003) 0.12555	0.006 (−0.007, 0.020) 0.34420	−0.000 (−0.014, 0.013) 0.99059
Female	0.016 (0.000, 0.032) 0.04795*	0.014 (0.001, 0.027) 0.04174*	0.007 (−0.006, 0.020) 0.28035
Subgroup analysis, stratified by race/ethnicity
Mexican American	0.039 (0.007, 0.072) 0.01866*	0.018 (−0.010, 0.047) 0.20766	0.013 (−0.015, 0.042) 0.35671
Other Hispanic	0.009 (−0.028, 0.046) 0.64239	−0.007 (−0.039, 0.025) 0.67295	−0.013 (−0.046, 0.019) 0.41541
Non-Hispanic White	−0.001 (−0.017, 0.015) 0.88818	0.008 (−0.006, 0.022) 0.25430	0.001 (−0.013, 0.015) 0.90416
Non-Hispanic Black	0.007 (−0.020, 0.034) 0.61000	0.014 (−0.010, 0.038) 0.24613	0.004 (−0.020, 0.029) 0.73143
Other race	0.049 (0.026, 0.073) 0.00004 †	0.016 (−0.004, 0.037) 0.11903	0.011 (−0.010, 0.031) 0.30919

Model 1: Non-adjusted; Model 2: adjusted for sex and age; Model 3: adjusted for sex, age, race, marital status, total number of people in the household, smoking status, alcohol use, TG, TC, serum creatinine, uric acid, and plasma glucose.

*P < 0.05, ^†P < 0.01.

Table 6

Association between Serum Selenium (μg/L) and Diastolic Blood Pressure (mmHg).

	Model 1 β/OR (95%CI) P value	Model 2 β/OR (95%CI) P value	Model 3 β/OR (95%CI) P value
Serum selenium (μg/L)	0.036 (0.029, 0.043) <0.00001 †	0.034 (0.027, 0.041) <0.00001 †	0.024 (0.017, 0.031) <0.00001 †
Subgroup analysis, stratified by sex
Male	0.037 (0.027, 0.047) <0.00001 †	0.038 (0.027, 0.048) <0.00001 †	0.026 (0.016, 0.037) <0.00001 †
Female	0.030 (0.020, 0.039) <0.00001 †	0.030 (0.020, 0.039) <0.00001 †	0.022 (0.013, 0.032) <0.00001 †
Subgroup analysis, stratified by race/ethnicity
Mexican American	0.043 (0.021, 0.064) 0.00012 †	0.034 (0.012, 0.055) 0.00230 †	0.027 (0.006, 0.049) 0.01157 †
Other Hispanic	0.044 (0.021, 0.067) 0.00020 †	0.037 (0.015, 0.060) 0.00134 †	0.021 (−0.002, 0.043) 0.07556 †
Non-Hispanic White	0.035 (0.024, 0.045) <0.00001 †	0.030 (0.020, 0.041) <0.00001 †	0.018 (0.007, 0.028) 0.00082 †
Non-Hispanic Black	0.039 (0.021, 0.057) 0.00002 †	0.038 (0.020, 0.056) 0.00003 †	0.029 (0.011, 0.047) 0.00137 †
Other race	0.036 (0.021, 0.052) <0.00001 †	0.029 (0.014, 0.044) 0.00017 †	0.025 (0.010, 0.040) 0.00096 †

Model 1: Non-adjusted; Model 2: adjusted for sex, age; Model 3: adjusted for sex, age, race, marital status, total number of people in the household, smoking status, alcohol use, TG, TC, serum creatinine, uric acid, and plasma glucose.

*P < 0.05, ^†P < 0.01.

In the subgroup analysis, serum Se levels were significantly associated with higher DBP (P for trend <0.001). However, this association did not reach statistical significance with serum Se levels (P = 0.566). In the adjusted model (adjusted for sex, age, and race), the association between serum Se levels and DBP was also significant (P < 0.001 for each trend). Participant data are summarized in Tables 7 and 8.

Table 7

ORs (95% CI) for Systolic Blood Pressure and Diastolic Blood Pressure by Quartiles of Serum Selenium Level in US Adults.

	Serum selenium (μg/L)	OR (95%CI)
Systolic pressure	Q1 (85.15–176.73)	Reference
	Q2 (176.73–191.325)	−0.628 (−1.461, 0.205)
	Q3 (191.325–207.108)	−0.707 (−1.541, 0.127)
	Q4 (207.108–734.8)	0.276 (−0.553, 1.105)
	P-trend	0.566
Diastolic pressure	Q1 (85.15–176.73)	Reference
	Q2 (176.73–191.325)	1.421 (0.880, 1.962)
	Q3 (191.325–207.108)	2.022 (1.480, 2.564)
	Q4 (207.108–734.8)	2.762 (2.223, 3.300)
	P-trend	<0.001 †

*P < 0.05, ^†P < 0.01.

Table 8

Adjusted ORs (95% CI) for Systolic Blood Pressure and Diastolic Blood Pressure by Quartiles of Serum Selenium Level in US Adults.

	Serum selenium (μg/L)	OR (95%CI)
Systolic pressure	Q1 (85.15–176.73)	Reference
	Q2 (176.73–191.325)	−0.008 (−0.729, 0.713)
	Q3 (191.325–207.108)	−0.002 (−0.728, 0.725)
	Q4 (207.108–734.8)	0.407 (−0.327, 1.142)
	P-trend	0.004 †
Diastolic pressure	Q1 (85.15–176.73)	Reference
	Q2 (176.73–191.325)	0.957 (0.435, 1.480)
	Q3 (191.325–207.108)	1.459 (0.932, 1.986)
	Q4 (207.108–734.8)	1.772 (1.240, 2.304)
	P-trend	<0.001 †

Adjusted for sex, age, and race.

*P < 0.05, ^†P < 0.01.

Discussion

This study investigated the correlation between serum Se levels and blood pressure. In selected representative individuals from a cross-sectional study of the US population enrolled in NHANES 2009–2018, participants with hypertension had higher serum Se levels than participants with normal blood pressure. Serum Se was significantly associated with DBP in both men and women after adjustment for many factors. Muhammad Imran, in his book titled “Distribution of Selenium in Soil and Human Health,” has indicated that Se in soil plays an important role in the food chain. The total Se content of the world’s soils ranges from 0.125 to 0.3 mg kg⁻¹, and varies with soil properties. High and low Se in soils can also pose severe environmental and health hazards. A major strength of this study was the inclusion of many populations and the ability to perform predefined subgroup analyses by sex and race/ethnicity. Differences in serum Se levels by race/ethnicity may be associated with lifestyle factors such as diet, location, and genetics, as shown in a review by Papp et al. [22–25].

Various studies have drawn differing conclusions regarding the relationship between serum Se levels and blood pressure [6, 15, 18]. A meta-analysis investigating the cooperation between serum Se and hypertension has indicated that serum Se has a protective effect on hypertension [26]; case-control and cross-sectional studies have indicated a harmful relationship between serum Se and hypertension [27]; and other studies have found no association. The effect differs by sex and age, possibly because hypertension is closely associated with sex and age. Moreover, serum Se content increases with age, and the serum Se content is greater in men than in women of the same age [28, 29].

In addition, we used the 130/80 mmHg (SBP/DBP) criterion based on previous and current guidelines. Unknown confounders of hypertension might have changed over the years, thus leading to different findings regarding the relationship between serum Se levels and hypertension. Our study did not support a protective effect of serum Se on hypertension.

A 10-year prospective study has suggested a potentially detrimental association between Se levels and carotid artery wall thickness [30]. McDonald et al. have reported that patients with lower preoperative Se levels may have higher risk of developing postoperative atrial fibrillation following coronary artery bypass grafting [31]. Results from a cross-sectional study have indicated a direct relationship between serum Se and blood pressure in a high Se environment, and generally stronger associations in women than men [32]. A meta-analysis published in 2014 has indicated that no conclusive evidence supports an association between Se and hypertension [6]. A cross-sectional study investigating the relationship between Se and hypertension has suggested that high Se levels in women are more likely to increase the risk of developing hypertension than similar levels in men (OR 1.192 vs. 1.043) [27]. Findings from a study evaluating the role of Se in hypertension have demonstrated a significant positive correlation between serum Se levels and hypertension: serum Se levels ≥120 μg/L were significantly associated with hypertension after adjustment for confounders (OR 1.46). The correlation with the development of hypertension increased at serum Se levels >150 μg/L (OR 1.69) [18].

Recent epidemiological studies have reported that high Se levels may contribute to high blood pressure [33]. High serum Se levels are associated with a high prevalence of hypertension in US adults [15]. Our study indicated a potential sex difference in the association between serum Se levels and blood pressure.

Additionally, we fully adjusted for covariates in our analysis, including indicators of socioeconomic status. Furthermore, we used the most recent definition of hypertension, thus supporting the relevance of our findings to clinical practice. Nevertheless, the present study had several limitations. First, we used serial cross-sectional data, which cannot be used to determine a causal relationship between serum Se levels and hypertension. Second, if the definition in the most recent guidelines were used, the estimates of the prevalence of uncontrolled hypertension might have been higher.

Conclusions

In summary, the present analysis provides evidence supporting a nonlinear association between serum Se levels and the prevalence of hypertension. In particular, patients with hypertension exhibited higher serum Se levels than participants with normal blood pressure. The association between serum Se levels and DBP was also significant. However, this finding must be confirmed in future prospective cohort studies. If this association is confirmed, it will have substantial public health relevance. given the current and predicted future epidemic of hypertension.

Data Availability Statement

This study analyzed publicly available datasets. These data can be found at https://www.cdc.gov/nchs/nhanes/index.htm.

Ethics Statement

The study involving human participants was reviewed and approved by the NCHS Research Ethics Review Board. Written informed consent for participation was not required for this study, in accordance with the national legislation and the institutional requirements.

Author Contributions

Wang Lili: conceptualization, research design, writing, and editing; Guo Liwei: conceptualization, research design, and implementation. All authors have read and agreed to the published version of the manuscript.

Funding and Acknowledgments

This study received no financial support. The authors thank the NHANES study staff and participants for their valuable contributions.

Conflicts of Interest

The authors have no conflicts of interest to declare.

Citation Information

Download Citation.

References

1. KT Mills, A Stefanescu, J He. The global epidemiology of hypertension. Nat Rev Nephrol 2020;16(4):223–37.

2. RS Vasan. High blood pressure in young adulthood and risk of premature cardiovascular disease: calibrating treatment benefits to potential harm. J Am Med Assoc 2018;320(17):1760–63.

3. FD Fuchs, PK Whelton. High blood pressure and cardiovascular disease. Hypertension 2019;75(2):285–92.

4. D Di Raimondo, S Buscemi, G Musiari, G Rizzo, E Pirera, D Corleo, et al. Ketogenic diet, physical activity, and hypertension-A narrative review. Nutrients 2021;13(8):2567.

5. X Chen, H Jin, D Wang, J Liu, Y Qin, Y Zhang, et al. Serum creatinine levels, traditional cardiovascular risk factors and 10-year cardiovascular risk in Chinese patients with hypertension. Front Endocrinol (Lausanne) 2023;14:1140093.

6. D Kuruppu, HC Hendrie, L Yang, S Gao. Selenium levels and hypertension: a systematic review of the literature. Public Health Nutr 2014;17(6):1342–52.

7. JC Avery, PR Hoffmann. Selenium, selenoproteins, and immunity. Nutrients 2018;10(9):1203.

8. V Branco, J Canário, J Lu, A Holmgren, C Carvalho. Mercury and selenium interaction in vivo: effects on thioredoxin reductase and glutathione peroxidase. Free Radic Biol Med 2012;52(4):781–93.

9. A Razaghi, M Poorebrahim, D Sarhan, M Björnstedt. Selenium stimulates the antitumour immunity: insights to future research. Eur J Cancer 2021;155:256–67.

10. D Fernández-Lázaro, CI Fernandez-Lazaro, J Mielgo-Ayuso, LJ Navascués, A Córdova Martínez, J Seco-Calvo. The role of selenium mineral trace element in exercise: antioxidant defense system, muscle performance, hormone response, and athletic performance. a systematic review. Nutrients 2020;12(6):1790.

11. X Wang, YA Seo, SK Park. Serum selenium and non-alcoholic fatty liver disease (NAFLD) in U.S. adults: National Health and Nutrition Examination Survey (NHANES) 2011–2016. Environ Res 2021;197:111190.

12. H Steinbrenner, LH Duntas, MP Rayman. The role of selenium in type-2 diabetes mellitus and its metabolic comorbidities. Redox Biol 2022;50:102236.

13. MC Houston. The role of cellular micronutrient analysis, nutraceuticals, vitamins, antioxidants and minerals in the prevention and treatment of hypertension and cardiovascular disease. Ther Adv Cardiovasc Dis 2010;4(3):165–83.

14. L Lei, F Zhang, J Huang, X Yang, X Zhou, H Yan, et al. Selenium deficiency causes hypertension by increasing renal AT₁ receptor expression via GPx1/H₂O₂/NF-κB pathway. Free Radic Biol Med 2023;200:59–72.

15. M Laclaustra, A Navas-Acien, S Stranges, JM Ordovas, E Guallar. Serum selenium concentrations and hypertension in the US Population. Circ Cardiovasc Qual Outcomes 2009;2(4):369–76.

16. A Kuria, H Tian, M Li, Y Wang, JO Aaseth, J Zang, et al. Selenium status in the body and cardiovascular disease: a systematic review and meta-analysis. Crit Rev Food Sci Nutr 2021;61(21):3616–25.

17. Z Yuan, X Xu, H Ye, L Jin, X Zhang, Y Zhu. High levels of plasma selenium are associated with metabolic syndrome and elevated fasting plasma glucose in a Chinese population: A case-control study. J Trace Elem Med Biol 2015;32:189–94.

18. MM Bastola, C Locatis, R Maisiak, P Fontelo. Selenium, copper, zinc and hypertension: an analysis of the National Health and Nutrition Examination Survey (2011–2016). BMC Cardiovasc Disord 2020;20(1):45.

19. TS Nawrot, JA Staessen, HA Roels, E Den Hond, L Thijs, RH Fagard, et al. Blood pressure and blood selenium: a cross-sectional and longitudinal population study. Eur Heart J 2007;28(5):628–33.

20. LR Curtin, LK Mohadjer, SM Dohrmann, JM Montaquila, D Kruszan-Moran, LB Mirel, et al. The national health and nutrition examination survey: sample design, 1999–2006. Vital Health Stat 2 2012;(155):1–39.

21. JM Flack, B Adekola. Blood pressure and the new ACC/AHA hypertension guidelines. Trends Cardiovasc Med 2020;30(3):160–4.

22. M Imran, Z Chen, A Mehmood, S Rukh, W Weixie, W Asghar, et al. Distribution of Selenium in Soils and human Health. In: V Gelen, A Kara, A Kükürt, editors. 2023. https://www.intechopen.com/chapters/86999 .

23. LV Papp, A Holmgren, KK Khanna. Selenium and selenoproteins in health and disease. Antioxid Redox Signal 2010;12(7):793–5.

24. P Xun, D Bujnowski, K Liu, JS Morris, Z Guo, K He. Distribution of toenail selenium levels in young adult Caucasians and African Americans in the United States: the CARDIA Trace Element Study. Environ Res 2011;111(4):514–9.

25. CA Reyes-Ortiz, H Ju, K Eschbach, YF Kuo, JS Goodwin. Neighbourhood ethnic composition and diet among Mexican-Americans. Public Health Nutr 2009;12(12):2293–301.

26. P Agasthi, S Aloor, V Yarlagadda, A Chenna, A Onwuanyi. Association between circulating selenium levels and primary hypertension: a meta-analysis. Hypertension 2015;66:P050.

27. G Wu, Z Li, W Ju, X Yang, X Fu, X Gao. Cross-sectional study: relationship between serum selenium and hypertension in the Shandong Province of China. Biol Trace Elem Res 2018;185(2):295–301.

28. L Schomburg. Dietary selenium and human health. Nutrients 2016;9(1):22.

29. MP Rayman. Selenium intake, status, and health: a complex relationship. Hormones (Athens) 2020;19(1):9–14.

30. R Swart, AE Schutte, JM van Rooyen, CMC Mels. Selenium and large artery structure and function: a 10-year prospective study. Eur J Nutr 2019;58(8):3313–23.

31. C McDonald, J Fraser, K Shekar, A Clarke, J Coombes, A Barnett, et al. Low preoperative selenium is associated with post-operative atrial fibrillation in patients having intermediate-risk coronary artery surgery. Eur J Clin Nutr 2016;70:1138–43.

32. M Vinceti, R Chawla, T Filippini, C Dutt, S Cilloni, R Loomba, et al. Blood pressure levels and hypertension prevalence in a high selenium environment: results from a cross-sectional study. Nutr Metab Cardiovasc Dis 2019;29:398–408.

33. M Sobel, A Navas-Acien, M Powers, M Grau-Perez, W Goessler, LG Best, et al. Environmental-level exposure to metals and metal-mixtures associated with spirometry-defined lung disease in American Indian adults: Evidence from the Strong Heart Study. Environ Res 2022;207:112194.