Selenium Evidence

The Selenium–Vitamin D Axis: An Evidence-Graded Synthesis

Question: What is the association of selenium with vitamin D?

Claude AI - May 2026

TL;DR

Mechanistically plausible, observationally consistent, but clinically thin: Selenium and vitamin D share biological roles in oxidative stress, immune modulation, and thyroid biology, and observational data show modest positive correlations between serum selenium and 25(OH)D — but high-quality randomized evidence that combining the two yields outcomes superior to either alone exists in only a narrow slice of disease (chiefly Hashimoto's and Graves' disease) and even there derives from small, mostly single-center trials.
The strongest combined-supplementation signal is in autoimmune thyroid disease: Krysiak et al. (Pharmacol Rep 2019; n=47) showed that adding selenomethionine 200 μg/day to vitamin D 4000 IU/day for 6 months produced "more pronounced" reductions in TPOAb/TgAb than vitamin D alone, and the SeMMIviD randomized trial (Gallo et al., Front Endocrinol 2022; n=42) demonstrated that selenium + cholecalciferol added to methimazole achieved a 12.2 pg/mL greater drop in FT4 at 180 days (p=0.002) in newly diagnosed Graves' disease.
Outside thyroid autoimmunity the evidence is preliminary or null: cancer, cardiovascular, diabetes, PCOS, COVID-19, and reproductive-health data treat the two nutrients side-by-side rather than synergistically; mechanistic links to VDR, CYP enzymes, and DBP/GC are largely indirect and have not been demonstrated in human interventional studies.

Key Findings

Mechanistic links (Grade C — biologically plausible, mostly indirect). Selenium exerts most of its biology through ~25 selenoproteins, the most relevant of which to the vitamin D axis are the glutathione peroxidases (GPX1–4), thioredoxin reductases (TXNRD1–3), the three iodothyronine deiodinases (DIO1–3), and selenoprotein P (SELENOP, the major selenium transport protein). All three deiodinases contain selenocysteine in their active sites and regulate intracellular T3; their strongest point of contact with vitamin D is shared regulation of cellular redox tone and immune-cell function rather than direct biochemical coupling. No human study has demonstrated that selenium status modulates CYP2R1 (25-hydroxylation), CYP27B1 (1α-hydroxylation), or CYP24A1 (24-hydroxylation) activity; mechanistic claims about selenium altering vitamin D activation rest on indirect redox arguments. Vitamin D binding protein (DBP/GC) is structurally a member of the albumin gene family (along with albumin, α-fetoprotein, and afamin) but is not selenium-dependent; GC polymorphisms (rs2282679, rs4588, rs7041) influence circulating 25(OH)D independently of selenium status.

Shared redox/immune pathways. GPX1 expression is induced by vitamin D in human supplementation studies in Arab adults with prediabetes (Alkharfy et al., Nutrients 2020), providing rare direct evidence that vitamin D supplementation upregulates a selenoenzyme. Conversely, glutathione precursor supplementation has been shown to raise 25(OH)D levels and upregulate hepatic vitamin D regulatory genes (Jain et al., 2018; PMC6208166), suggesting a bidirectional redox–vitamin D loop in which selenium-dependent GPX activity participates.

Observational/epidemiological associations (Grade B — consistent but modest).

Wuhan birth cohort (Gang et al., Nutrients 2022;14(9):1715, n=1,695 mother–infant pairs): Each doubling of whole-pregnancy urinary selenium was associated with a 21.14% increase in cord 25(OH)D (95% CI 8.69–35.02%) after multivariable adjustment; trimester-specific increases ranged 8.76%–15.44%. Mothers in the lowest and middle selenium tertiles had higher odds of newborn vitamin D deficiency than those in the highest tertile.
NHANES 2011–2018 (Yi et al., BMC Public Health 2024, n=9,377): Serum selenium had a positive linear association with serum 25(OH)D after adjustment for confounders, contrasting with negative associations seen with cadmium, lead, and manganese; the selenium–CVD relationship was U-shaped.
NHANES 2011–2016 mediation analysis (Biomed Environ Sci 2024, n=9,146): Serum selenium was positively associated with both 25(OH)D and type-2 diabetes prevalence; vitamin D partially mediated the relationship between selenium and dysglycemia.
South Korean IBM Cohort Study (Lee et al., Eur Thyroid J. 2024;13(4):ETJ-24-0007, n=367 pregnant women): 8% selenium-deficient (<70 μg/L) and 45% suboptimal (70–99 μg/L); selenium deficiency was independently associated with positive anti-TPO antibodies (13.3% vs 4.6% in optimal, p=0.031) and ultrasound features of autoimmune thyroiditis.
Geographic overlap: Selenium deficiency clusters in parts of China (Keshan-disease belt), sub-Saharan Africa, and parts of Europe (Balkans, Scandinavia); suboptimal selenium and vitamin D each affect ~50% of the Serbian population (Stevanović et al., Front Nutr 2021). The overlap is partial — vitamin D deficiency is largely latitude- and skin-pigmentation-driven, whereas selenium status tracks soil composition and seafood/Brazil-nut intake.

Clinical trial data (Grade B for Hashimoto's/Graves'; Grade D elsewhere).

Hashimoto's thyroiditis — the most-studied combination.

Krysiak et al., Pharmacol Rep 2019;71(2):367–373 (n=47 euthyroid women, low vitamin D status): All received cholecalciferol 4000 IU/day for 6 months; 23 had been receiving selenomethionine 200 μg/day for ≥12 months prior and continued it. In both groups, 25(OH)D rose and TPOAb/TgAb fell, and SPINA-GT (thyroid secretory capacity) rose; "the effects on antibody titers and the SPINA-GT index were more pronounced in women receiving selenomethionine." The authors concluded: "selenium intake enhances the effect of vitamin D on thyroid autoimmunity… euthyroid women with thyroid autoimmunity may benefit more from vitamin D/selenomethionine combination therapy than from treatment with only one of these agents." Limitations: non-randomized allocation (based on prior selenium use), small sample, single center, no open-access numerical results table.
Krysiak et al., Pharmacol Rep 2019;71(2):243–247 (men, n=37, pilot): vitamin D 4000 IU/day vs selenomethionine 200 μg/day monotherapy for 6 months. Both reduced TPOAb (~26.4% and ~26.1%) and TgAb (~25.7% and ~27.2%) and increased SPINA-GT ~10–11%. Only selenium affected the SPINA-GD index; only vitamin D raised 25(OH)D.
Huwiler et al., Thyroid 2024 (systematic review/meta-analysis of selenium RCTs in HT): selenium reduced TSH in untreated patients and TPOAb levels; effects on TgAb, FT4, T3, and clinical outcomes were inconsistent. Vitamin D was not the focus, but the review highlighted considerable methodological heterogeneity. A separate 2025 meta-analysis (Medicine, PMC12401265, 21 studies, n=1,610) reported TPOAb standardized mean difference of −0.46 (95% CI −0.74 to −0.18, p=0.001) at 3 months and −0.80 (95% CI −1.38 to −0.21, p=0.008) at 6 months with selenium.
Peng et al., Front Endocrinol 2024 (network meta-analysis of supplements in HT): combinations of vitamin D + selenium (± inositol) produced the largest TPOAb reductions versus single agents, but evidence certainty was low.
Feng et al. retrospective study in T2DM + HT (n=150, Central Hospital of Shaoyang, China): adding vitamin D 4000 IU + selenium 200 μg/day to standard antidiabetic therapy for 3–6 months improved thyroid function, TPOAb, glycemia, and lipids compared to antidiabetic drugs alone.
Graves' disease — SeMMIviD (Gallo D et al., Front Endocrinol 2022;13:886451; EudraCT 2017-005050-11):
Randomized controlled trial, n=42 newly diagnosed Graves' patients with baseline selenium <120 μg/L and 25(OH)D <30 ng/mL. Methimazole alone vs methimazole + selenium 100 μg/day (selenomethionine 83 μg + Se yeast 17 μg) for 6 months + cholecalciferol (loading bolus then 7,000 IU/week ≈1,000 IU/day) for 9 months. Verbatim primary result:
"Combination therapy resulted in a significantly greater reduction in serum FT4 concentration at 45 days (−37.9 pg/ml, CI 95%, −43.7 to −32.2 pg/ml) and 180 days (−36.5 pg/ml, CI 95%, −42 to −30.9 pg/ml) compared to MMI monotherapy (respectively: −25.7 pg/ml, CI 95%, −31.6 to −19.7 pg/ml and −22.9 pg/ml, CI 95%, −28 to −17.3 pg/ml, p 0.002)."
Between-arm mean difference in FT4 variation was 12.2 pg/mL (p=0.002). ThyPRO quality-of-life improvement was significantly greater in the combination arm at all time points (−14.3 vs −3.5 at 270 days, p=0.003). TRAb and FT3 between-arm differences were not statistically significant. A follow-on analysis (Front Endocrinol 2023) reported increased regulatory-T-cell frequency and altered NK-cell phenotype in the intervention arm.
COVID-19 (Grade D — observational, vulnerable to confounding).
Heller et al. (Redox Biol 2021) and others documented strikingly lower serum selenium and 25(OH)D in COVID-19 fatalities than survivors; a small cohort of Hashimoto's patients (Hassan et al., ECE 2021; n=356) reported that patients pre-taking selenium + zinc + vitamin D had milder COVID-19 and lower hospitalization rates (28% hospitalization in non-supplemented; significantly lower in supplemented, p<0.05). No randomized trial has specifically tested combined selenium + vitamin D supplementation for COVID-19 outcomes.
Cancer (Grade D for the combination).
Both micronutrients have been studied separately in colorectal, prostate, lung, and bladder cancer. The Nutritional Prevention of Cancer (NPC) trial (1,312 participants, 200 μg/day selenium) suggested reduced overall cancer incidence among baseline-deficient participants, but the Selenium and Vitamin E Cancer Prevention Trial (SELECT) found no benefit and a possible prostate-cancer harm signal. The Cochrane review (Vinceti et al., Cochrane Database Syst Rev 2018;1:CD005195) concluded no protective effect of selenium against cancer in RCTs. The VITAL trial (Manson JE et al., N Engl J Med. 2019;380:33–44) showed that *"daily high-dose vitamin D supplementation for 5 years among initially healthy adults did not reduce incidence of cancer"
— HR 0.96 (95% CI 0.88–1.06) for total invasive cancer incidence. No major trial has tested the selenium + vitamin D combination as a co-intervention for cancer prevention.
Cardiovascular disease (Grade C).
Narayanam H, Chinni SV, Samuggam S. "The Impact of Micronutrients-Calcium, Vitamin D, Selenium, Zinc in Cardiovascular Health: A Mini Review." Front Physiol. 2021;12:742425 concluded that adequate but not excessive levels of both nutrients support cardiovascular health, with U-shaped risk curves and supplementation trials showing little benefit at population level. NHANES mediation analyses suggest 25(OH)D can mediate part of the selenium–CVD relationship. A separate NHANES 2011–2016 analysis (Front Med. 2025;PMC12006169, n=645 adults) found a selenium threshold of 106.8 μg/L for diabetic retinopathy risk (OR = 0.88, p = 0.0107).

Metabolic syndrome/diabetes (Grade C — conflicting).

The NHANES 2003–2004 analysis (Laclaustra M et al., Environ Health Perspect 2009;117:1409–1413) found an OR of 7.64 (95% CI 3.34–17.46) for diabetes in the highest selenium quartile (≥147 μg/L) vs the lowest (<124 μg/L). A pooled summary from a meta-analysis of 13 observational studies (Kohler LN et al., Nutrients 2018;10:1924) gave OR 2.03 (95% CI 1.51–2.72), whereas RCTs of selenium have not shown increased diabetes risk (OR 1.18, 0.95–1.47; Kohler et al.). Vitamin D RCTs (D2d, VITAL) have failed to show reduced type-2 diabetes incidence in non-deficient adults. The two nutrients have been studied jointly mainly in PCOS, gestational diabetes, and Hashimoto's-plus-T2DM, where signals are present but driven largely by retracted or expression-of-concern Asemi-group publications.

PCOS (Grade D — heterogeneous, with signals of concern). Selenium 200 μg/day for 12 weeks worsened insulin resistance in one PCOS trial (HOMA-IR 2.05 vs 1.81, p=0.017; Mehdi et al. 2016), while other Asemi-group trials reported metabolic benefit; several of those publications now carry expressions of concern. Vitamin D supplementation in PCOS produces marginal metabolic improvements in deficient women (Trummer et al., Eur J Nutr 2019). The combination has not been tested in well-powered randomized trials.

Premature ovarian insufficiency (POI) and ovarian reserve (Grade D). A 2025 systematic review (Pargar et al., Health Sci Rep, n=12 studies, 521 women) found associations between low serum vitamin D, vitamin E, and selenium and POI/ovarian reserve markers, but results were inconsistent across studies and the body of evidence comprised mostly observational designs.

Pregnancy outcomes (Grade C). The Wuhan cohort showed a positive selenium–25(OH)D association (above); the Lee et al., Eur Thyroid J. 2024 IBM Cohort Study tied selenium deficiency to severe preeclampsia in twin pregnancies (0% in selenium-supplemented vs 9.0% in non-supplemented, p=0.015). Vitamin D supplementation in GDM (Iranian Asemi/Karamali trials) reduced neonatal hospitalization (RR 0.40, 95% CI 0.23–0.69), but these data are subject to the same publication-integrity concerns.

Disease contexts and effect-size summary. The selenium–vitamin D axis has been most rigorously studied in autoimmune thyroid disease; the data support combined supplementation when patients are deficient in both nutrients. Outside thyroid disease, mechanistic plausibility consistently outpaces clinical proof.

Details

Mechanistic biochemistry. Selenium is incorporated as the 21st amino acid selenocysteine into ~25 human selenoproteins. The selenoproteins most relevant to vitamin D biology are:- GPX1, GPX3, GPX4: cytosolic, plasma, and membrane glutathione peroxidases that reduce hydrogen peroxide and lipid hydroperoxides; protect VDR-expressing immune cells (T cells, dendritic cells, macrophages) from oxidative damage.- TXNRD1–3: regenerate reduced thioredoxin, supporting ribonucleotide reductase, DNA repair, and redox signaling — pathways that intersect vitamin D's anti-cancer effects.- DIO1–3: the iodothyronine deiodinases, selenocysteine-active-site enzymes that activate/inactivate thyroid hormones. Vitamin D and thyroid hormone receptors share retinoid-X-receptor (RXR) heterodimerization partners, suggesting cross-talk at the transcription-factor level.- SELENOP: the principal Se transport protein (carries >50% of plasma selenium); falling SELENOP is the most sensitive biomarker of inadequate selenium status. SELENOP shares no structural homology with DBP/GC, but both function as plasma carrier proteins with target-tissue uptake mediated in part by megalin (megalin handles both SELENOP and DBP–25(OH)D complexes — a potential point of indirect functional intersection at the renal proximal tubule).

Vitamin D pathway components and where selenium could (but has not been shown to) intersect:

25-hydroxylation (CYP2R1, hepatic): No direct evidence selenium status modulates CYP2R1; selenium deficiency–associated hepatic oxidative stress could in principle impair microsomal CYP function, but this is theoretical.
1α-hydroxylation (CYP27B1, renal and extrarenal): Activated in macrophages and dendritic cells by IFN-γ and PRR signaling; selenium-dependent redox tone influences macrophage activation, providing an indirect link.
24-hydroxylation/inactivation (CYP24A1): Induced by 1,25(OH)2D as feedback; no selenium dependence shown.
VDR signaling: VDR/RXR heterodimers bind VDREs at thousands of target genes; selenium has not been shown to alter VDR ligand binding or DNA-binding affinity, but oxidative stress can inhibit nuclear-receptor coactivator function.
DBP/GC: A multifunctional protein in the albumin family; total DBP levels are not influenced by vitamin D supplementation per se (Sollid et al., BMC Endocr Disord 2016), and selenium has not been shown to modulate DBP synthesis, glycosylation, or actin-scavenging function.

Observational data in detail. The strongest single dataset is the Wuhan birth cohort (Gang et al.): "each doubling of urinary Se concentrations in the first, second, third trimester, and whole pregnancy (average SG-adjusted concentrations across three trimesters) were associated with 8.76% (95% CI: 4.30%, 13.41%), 15.44% (95% CI: 9.18%, 22.06%), 11.84% (95% CI: 6.09%, 17.89%), and 21.14% (95% CI: 8.69%, 35.02%) increases in 25(OH)D levels." NHANES analyses uniformly confirm a positive selenium–25(OH)D correlation in U.S. adults, though both nutrients independently associate with diet quality, socioeconomic status, and supplement use — residual confounding cannot be excluded.

The two pivotal combination RCTs in detail.

SeMMIviD (Gallo et al., 2022; EudraCT 2017-005050-11): 42 Caucasian patients with newly diagnosed Graves' disease and baseline Se <120 μg/L plus 25(OH)D <30 ng/mL randomized 1:1 to methimazole alone vs methimazole + selenium 100 μg/day (selenomethionine + selenium yeast) for 6 months plus vitamin D (loading bolus then 7,000 IU/week for 9 months). The combination arm achieved a between-arm FT4 reduction of 12.2 pg/mL greater than control (p=0.002), with mean FT4 normalizing at 45 days in the intervention arm. ThyPRO quality-of-life improvement was significantly greater at all time points (e.g., −14.3 vs −3.5 at 270 days, p=0.003). No selenosis or hypercalcemia. The authors concluded: "reaching optimal Se and VitD levels increases the early efficacy of MMI treatment when Se and VitD levels are suboptimal." Limitations: single center, single-blinded (open MMI), early termination at n=42 due to the pandemic, baseline disease severity imbalance favoring control.

Krysiak combination studies (multiple, 2018–2023): The same Polish group has published more than ten small studies in Hashimoto's combining vitamin D and selenomethionine, generally finding that the combination outperforms either alone for antibody reduction. The largest single combination study (Pharmacol Rep 2019, n=47) reported the qualitative finding that combination was "more pronounced" without providing percent-reduction tables in open-access form. Caveats: all from one center, mostly euthyroid women with low vitamin D, non-randomized allocation in several reports, no hard clinical endpoints (TSH normalization, levothyroxine sparing, disease progression).

Evidence Quality Assessment

Strong findings (Grade A/B):- Selenium is essential for deiodinase activity and thyroid hormone homeostasis (Grade A).- Selenium supplementation reduces TPOAb in Hashimoto's patients short-term (Grade B, multiple meta-analyses).- Vitamin D and selenium are positively correlated in serum in multiple large observational cohorts (Grade B).- Combined selenium + vitamin D added to methimazole speeds biochemical control of Graves' hyperthyroidism (Grade B, one well-conducted RCT — SeMMIviD).

Weaker/preliminary (Grade C/D):

"Synergy" of selenium and vitamin D in Hashimoto's (Grade C — based on small, mostly single-group studies; the Krysiak series is suggestive but not definitive).
Combined supplementation for COVID-19, cancer, cardiovascular, PCOS, GDM, POI — all Grade D (preliminary, observational, or contaminated by trial-integrity concerns).
Direct selenium → CYP2R1/CYP27B1/CYP24A1 modulation: not established in humans (Grade D, mechanistic speculation).

Major research gaps:

No large multicenter RCT of vitamin D ± selenium with co-primary endpoints in any single disease.
No human data on whether selenium status modulates CYP27B1/CYP24A1 activity or VDR transcriptional output.
Lack of pre-specified subgroup analyses by baseline selenium and 25(OH)D status in existing vitamin D trials (VITAL, D2d) — the VITAL Hahn et al. 2022 autoimmune-disease substudy did not stratify by selenium.
Trial-integrity concerns: Many of the most-cited GDM/PCOS combination trials originate from a single Iranian research group; several have been retracted or carry expressions of concern.
Mechanistic disconnect: although both nutrients influence T-helper-cell balance and oxidative stress, no human study has measured selenoprotein P, GPX activity, and 1,25(OH)2D simultaneously to test the redox-mediation hypothesis.

Important caveats:- Both nutrients exhibit U-shaped risk curves: serum selenium >~140 μg/L correlates with increased type-2 diabetes incidence (Laclaustra 2009 highest-quartile OR 7.64); vitamin D >125 nmol/L correlates with adverse outcomes in some NHANES analyses. Combination supplementation should target physiologic replacement, not maximization.- Geographic context matters: in selenium-replete populations (USA, Venezuela), selenium supplementation may confer no benefit and possible harm; in deficient populations (parts of Europe, China, sub-Saharan Africa) it is more likely beneficial.- Most positive trials are in deficient populations; null trials in replete populations do not refute deficiency-correction benefit.

Recommendations

For clinicians caring for patients with autoimmune thyroid disease:1. Measure both 25(OH)D and selenium status (serum selenium ± SELENOP) at baseline in newly diagnosed Hashimoto's or Graves' disease. This is the clinical context where combined repletion has the best evidence base.2. If both are insufficient (25(OH)D <30 ng/mL and serum Se <120 μg/L), consider repletion to physiologic ranges: cholecalciferol 1,000–4,000 IU/day to achieve 25(OH)D 30–50 ng/mL; selenium 100–200 μg/day as selenomethionine or selenium yeast for 3–6 months, with reassessment. The SeMMIviD and Krysiak regimens are the most evidence-supported.3. Do not supplement selenium >200 μg/day or chronically in already-replete patients — U-shaped diabetes risk (NHANES highest-quartile OR 7.64) and observational signal for selenosis at chronic intakes >400 μg/day.

For clinicians caring for other patient populations:4. Do not routinely co-supplement selenium and vitamin D outside thyroid autoimmunity in the absence of documented deficiency. The cardiovascular, cancer, COVID-19, PCOS, and diabetes evidence does not currently justify combined empirical dosing.5. In pregnancy, ensure dietary or prenatal-vitamin sufficiency of both nutrients (vitamin D ≥600 IU/day; selenium ≥60 μg/day per IOM RDA); avoid high-dose supplementation without indication.

Thresholds that would change these recommendations:- A large multicenter RCT (≥500 participants) of vitamin D ± selenium in Hashimoto's with hard endpoints (progression to overt hypothyroidism, levothyroxine sparing) showing significant benefit would shift selenium from optional to standard adjunct.- A subgroup analysis of VITAL or D2d demonstrating selenium-status interaction would clarify whether selenium repletion meaningfully amplifies vitamin D's extra-skeletal benefits.- Confirmation of a U-shaped diabetes risk with combined supplementation would trigger explicit upper-limit guidance.

For researchers:6. Co-measure SELENOP, GPX activity, and 1,25(OH)2D in future vitamin D trials.7. Conduct factorial RCTs (vitamin D × selenium × placebo) in patients with documented dual insufficiency.8. Specifically test whether selenium status modulates VDR target-gene expression in vivo (e.g., cathelicidin, CYP24A1) in interventional designs.

Caveats

Many cited combination trials are small, single-center, and conducted by a small number of research groups; several Iranian PCOS/GDM publications are under expression of concern or have been retracted, materially limiting the cited evidence base.
The Krysiak Hashimoto's series, while suggestive, is non-randomized and concentrated at a single Polish center; the lack of open-access numerical effect tables limits independent evaluation.
Most observational selenium–25(OH)D correlations cannot exclude reverse causation or confounding by diet quality, supplement use, sun exposure, and socioeconomic status.
"Synergy" is a commonly used but rarely formally tested claim in this literature; only the Krysiak Pharmacol Rep 2019 women's study and the SeMMIviD Graves' trial directly compare combined versus single-nutrient interventions, and even these do not include a placebo-only arm.
Selenium toxicity (selenosis) is real and seen at chronic intakes >400 μg/day; physiologic plasma optimum (≈100–130 μg/L) and the U-shaped risk curve must be respected.
This synthesis does not constitute medical advice; individual decisions about supplementation require clinician evaluation including baseline nutritional status, comorbidities, and current medications.

Related in Vitamin D Life

Selenium is important to health - many studies