Cohort name	Original Cohort (N total)	Ascertainment of outcome	Exposure groups(a)	Incident cases	Model covariates	Results
Country	Exclusion criteria		n/person‐years
Reference	Study population (n, sex and age at baseline(a))
Follow‐up
Funding
BEST Bangladesh Bulka et al. (2019) 6 yr Prospective Cohort Public	N = 255 Population sampled: general population in an arsenic‐endemic area, randomly selected from placebo arm of BEST trial; adults having manifested arsenical skin lesions. Exclusion: pregnant women, unwillingness to discontinue vitamin use, prior history of cancer, too ill to participate, unwilling to give blood and urine samples; hypertensive at baseline. n = 178 Sex (% F): 54.9 Ethnicity: NR Age (yr): 24–64	SBP and DBP (2 seated measurements using an automated sphygmomanometer) measured in visits at 2, 4‐ and 6‐yr follow‐up. HTN defined as SBP ≥ 140 mmHg, DBP ≥ 90 mmHg, antihypertensive medication use, or self‐reported physician diagnosed HTN.	Blood Se (μg/L, median (IQR)): 121 (110.0–135.0) Cut‐offs for quartiles NR	Incident HTN cases: 46 N per quartile NR	Model 1 (M1): age and sex Model 2 (M2): M1 + baseline blood lead, manganese, and selenium Model 3 (M3): M2 + site, smoking status, educational duration, creatinine‐corrected urinary arsenic concentration Model 4 (M4): M3 + time‐varying BMI, time‐varying diabetes status	HR (95% CI) for incidence of HTN Q1(ref): 1 Q2: 1.31 (0.52, 3.29) Q3: 1.68 (0.70, 4.02) Q4: 1.33 (0.53, 3.36) p trend 0.432 Model 2 Q1(ref): 1 Q2: 1.17 (0.46, 3.01) Q3: 1.35 (0.55, 3.23) Q4: 1.11 (0.42, 2.91) p trend 0.593 Model 3 Q1(ref): 1 Q2: 0.94 (0.35, 2.54) Q3: 1.10 (0.41, 3.00) Q4: 0.90 (0.30, 2.68) p trend 0.882 Model 4 Q1(ref): 1 Q2: 0.72 (0.25, 2.08) Q3: 0.91 (0.32, 2.55) Q4: 0.73 (0.23, 2.30) p trend 0.641
FLEMENGHO Belgium Nawrot et al. (2007) 8.4 yr prospective cohort Public	N = 1,107 Population sampled: general population Exclusion: participants who died, became severely ill or moved; creatinine excretion was outside published limits; missing values; high BP or hypertension at baseline. n = 385 sex (% F): 51.8 Ethnicity: NR Age (yr) = 48.8 (> 20)	Blood pressure measured (5 seated measurements) by trained nurses, twice at baseline (1–3 wk apart) and once at follow‐up visit. High BP was defined as HTN (≥ 140/≥ 90 mmHg), high‐normal BP (130–139/85–89 mmHg), or self‐reported start of antihypertensive drug.	Blood Se (μg/L): 97.0 ± 19.0	Incident cases of high BP: 139; 36 cases per 1,000 person‐years	Age, BMI, smoking, the 24 h urinary excretion of sodium and potassium + for F: menopausal status and the use of contraceptive pills at baseline.	HR (95% CI) for incidence of high BP, by 20 μg/L blood Se at baseline F: 1.08 (0.90, 1.27) p = 0.41 M: 0.63 (0.48, 0.83) p = 0.0013
The Selenium and Cognitive Decline study China Su et al. (2016) 7 yr prospective cohort Public	N = 2000 Population sampled: general population > 65 yr Exclusion: severe hearing loss (cognitive assessment not possible). n = 635 sex (% F): 53.7 age (yr): 71.9 ± 5.6	Blood pressure measures and self‐reported HTN history collected at baseline, 2.5 and 7 yr follow‐up. HTN was defined as ≥ 140/≥ 90 mmHg.	Nail Se (μg/g): 0.413 ± 0.183 n, per quintile: Q1 ≤ 0.233: 402 Q2 0.234–0.362: 405 Q3 0.363–0.442: 395 Q4 0.443–0.552: 397 Q5 > 0.552: 401	Incident HTN cases n, per quintile: Q1: 192 Q2: 132 Q3: 120 Q4: 105 Q5: 86	Age, gender, BMI, education, smoking, alcohol consumption, physical activity.	HR (95% CI) for incidence of HTN Q1(ref): 1 Q2: 1.41 (1.03, 1.94) p = 0.0331 Q3: 1.93 (1.40, 2.67) p < 0.0001 Q4: 2.35 (1.69, 3.26) p < 0.0001 Q5: 1.94 (1.36, 22.77) p = 0.0002
PURE South Africa Swart et al. (2019) 10 yr prospective cohort Private	N = 987 Population sampled: general population Exclusion: missing baseline or follow‐up cardiovascular data or baseline Se data. n = 690 sex (% F): 64.6 Ethnicity: black South Africans Age (yr): 50.7 ± 10.2 (35–70)	Blood pressure (seated measurements; twice, 5‐min apart) measured with a validated device at baseline and follow‐up.	Plasma Se (μg/100 mL) Group with “normal Se status” at baseline (n = 845): 12.7 ± 3.41 Group with “deficient Se status” at baseline (n = 142): 6.12 ± 1.67	NA	Age, sex, BMI, physical activity index, tobacco use, Ɣ‐glutamyl transferase, glucose, C‐reactive protein, LDL	SBP level, β coefficients (95% CI), per 10 μg/L Se increase All: −0.03 (−0.10, 0.04) p = 0.418 Group with “normal Se status”: −0.03 (−0.11, 0.05) p = 0.439 Group with “deficient Se status”: 0.01 (−0.20, 0.22) p = 0.929 DBP NR
PROGRESS Mexico Kupsco et al. (2019) 6 yr prospective cohort Public	N = 948 Population sampled: mother–child pairs Exclusion: Women exposed to environmental tobacco smoke (at home) or smoking during pregnancy; missing values n = 548 sex (% F): 49.6 Ethnicity: NR Maternal age (yr) = 28 ± 5.6	Children blood pressure measures at 4–6 yr of age (4.8 ± 0.55 yr) during a clinical examination (seated measurements; twice, 3 min apart)	Maternal blood Se (2nd trimester) (μg/dL): 25 ± 4.5	NA	Single‐metal model (SMM): Maternal age, pre‐pregnancy BMI, education, socioeconomic status, parity (primiparous or multiparous), and environmental tobacco smoke (present or absent in home) Multi‐metal model (SMM): as above + As, Cd, Co, Cr, Cs, Cu, Mn, Pb, Sb, and Zn	BP levels, β‐coefficients (95% CI), per 1 μg/L Se increase SBP SMM: −0.19 (−0.51, 0.13) p = 0.25; MMM: −0.07 (−0.52, 0.39) p = 0.77 DBP SMM: β = −0.12 (−0.39, 0.14) p = 0.36; MMM: −0.14 (−0.51, 0.23) p = 0.46
RHEA Greece Howe et al. (2021) 11 yr prospective cohort Public	N = 1,363 Population sampled: mother–child pairs from Heraklion, Greece; aged ≥ 16 yr Exclusion: outliers for metal measurements (mean ± 4 SD), missing values n = 176 Sex (% F) = 44.3 Ethnicity: Caucasian Maternal age (yr) = 30.3 ± 4	Children blood pressure measured at 4, 6 and 11 yr of age during clinical examination (seated measurements; 3 times, 1‐min apart)	Maternal urinary Se in early pregnancy (specific‐gravity adjusted) (μg/L, geometric mean (95% CI)): 21.72 (20.72, 22.77)	Incident cases of elevated BP: 56	Maternal age, maternal education, maternal pre‐pregnancy BMI, maternal smoking during pregnancy, child's sex, child's exact age, and child's height at each time point, maternal urinary Co, Se, Mo, As, Cd, Sb, Pb	Per‐year change in BP levels from 4 to 11 yr of age, main effect estimate* for Se (95% CI) SBP β = 0.1 (−0.1, 0.3) DBP β = 0.1 (−0.1, 0.3) *Bayesian Varying Coefficient Kernel Machine Regression

BEST: The Bangladesh Vitamin E and Selenium Trial; BP: blood pressure; FLEMENGHO: The Flemish Study on Environment, Genes and Health Outcomes; HTN: Hypertension; PROGRESS: Programming Research in Obesity, Growth, Environment and Social Stressors; NR: not reported; PURE: Prospective Urban Rural Epidemiology; yr: year.

(a) Mean ± SD (range), unless specified otherwise.