TABLE 32

Key epidemiological studies on other effects and As exposure.

Reference, study population age, designOutcome definitionPopulation size (n) case/controlArsenic concentration/exposureResultsAdditional information/confounders
Mortality
Sohel et al. (2009), Bangladesh/> 15 years, cohort studyCancer, cardiovascular mortality, infectious, nonaccidental mortalityDeaths/survived: 9015/106888

w‐As (μg/L) (average in well water)

< 10

10–49

50–149

150–299

> 300

HR (95% CI) for cancer mortality:

1.00 (ref),

1.10 (0.77–1.59),

1.44 (1.06–1.95),

1.75 (1.28–2.40),

1.56 (1.06–2.30), p trend: 0.007.

HR (95% CI) for cardiovascular mortality:

1.00 (ref),

1.03 (0.82–1.29),

1.16 (0.96–1.40),

1.23 (1.01–1.51)

1.37 (1.07–1.77), p trend: 0.026.

HR (95% CI) for infectious mortality:

1.00 (ref),

1.09 (0.92–1.30)

1.30 (1.13–1.49),

1.51 (1.31–1.75),

1.59 (1.33–1.91), p trend: < 0.001.

HR (95% CI) for nonaccidental mortality:

1.00 (ref),

1.16 (1.06–1.26),

1.26 (1.18–1.36),

1.36 (1.27–1.47),

1.35 (1.23–1.48), p trend: < 0.001

Adjusted for age, sex, asset score, education
Wade et al. (2009), China/age data not reported, cross sectional studyAll‐cause mortality, cancer mortality, heart disease mortality, stroke mortality

Deaths/

survived: 572/12600

w‐As,

for 50 μg/L As increase

IRR (95% CI) and all‐cause mortality: 1.02 (0.94, 1.11), p: NS;

cancer mortality:

1.07 (0.89, 1.28), p: NS;

heart disease:

1.12 (1.01, 1.23), p: 0.034;

stroke:

0.82 (0.65, 1.03), p: NS

Adjusted for age, sex, education, smoking, drinking, farm work. Data also for different time of exposure
Argos et al. (2010), Bangladesh/18–75 years, cohort study (HEALS)All‐cause mortality, mortality due to chronic diseaseDeaths/survived: 407/11339

w‐As (μg/L)

0–10,

10.1–50,

50.1–150,

> 150

HR (95% CI) for all‐causemortality:

1.00 (ref),

1.34 (0·99–1·82),

1.09 (0·81–1·47),

1.68 (1·26–2·23).

HR (95%CI) for chronic‐diseasemortality:

1.00 (ref),

1.33 (0·94–1·87),

1.22 (0·87–1·70),

1.68 (1·21–2·33)

Adjusted for age, sex, body‐mass index, systolic blood pressure, education, smoking status. Data also for As dose and relation to change in u‐As over time
Rahman et al. (2013), Bangladesh/5–18 years, cohort (7‐years follow‐up)All‐cause mortality, cancer and cardiovascular mortalityDeaths/survived: 185/58221

Baseline w‐As (μg/L)

< 10

10–50

51–150

151–300

> 300

Cumulative As exposure (μg‐years/L)

< 1000

1000–4000

> 4000

HR (95% CI) for baseline w‐As and all‐cause mortality:

1.00 (ref),

1.13 (0.65–1.96),

0.81 (0.45–1.46),

1.35 (0.92–1.97),

1.51 (1.01–2.23), p trend < 0.05.

HR (95% CI) for cumulative w‐As and all‐cause mortality:

1.00 (ref),

1.17 (0.84–1.65),

1.90 (1.25–2.89), p trend < 0.05.

HR (95%CI) for baseline w‐As and cancer and cardiovascular mortality:

1.00 (ref),

1.53 (0.51–4.57),

1.29 (0.43–3.87),

2.18 (1.15–4.16), p trend < 0.05.

Adjusted for baseline age, educational attainment, socioeconomic status. Data also for time‐weighted lifetime average arsenic in well water.
Rahman et al. (2019), Bangladesh/5–18 years, cohort (13‐years follow‐up)Cancer and cerebrovascular‐cardiovascular‐respiratory mortality

Deaths/survived: cancer 48/58358;

Cerebro

vascular‐cardiovascular‐respiratory: 40/58366

Current w‐ As (μg/L)

< 90.9

90.9–223.1

> 223.1

cumulative w‐As exposure (μg‐years/L)

< 1013.3

1013.3–2711.0

> 2711

average or lifetime average As in well water (μg/L)

< 90.9

90.9–223.1

> 223.1

HR (95% CI) for average w‐As and cancer mortality:

1.00 (ref),

1.2 (0.6–2.3),

0.8 (0.4–1.6), p trend: NS.

HR (95% CI) for cumulative w‐As and cancer mortality:

1.00 (ref),

0.9 (0.5–1.8),

0.7 (0.3–1.5), p trend: NS.

HR (95%CI) for average w‐As and cardio‐cerebro‐respiratory mortality:

1.00 (ref),

2.7 (1.0–7.6),

4.8 (1.8–12.8), p trend: 0.01.

HR (95%CI) for cumulative water As and cardio‐cerebro‐respiratory mortality:

1.00 (ref),

4.0 (1.3–12.0),

5.1 (1.7–15.1), p trend: 0.01

Adjusted for baseline age, sex, educational attainment, socioeconomic status. Data also for time‐weighted lifetime average arsenic in well water.
Endocrine effects
Rahman, Kippler, et al. (2021), Bangladesh/1–15 years, birth cohort study (MINIMat/AsMat)Age at menarcheTotal 809

w‐As, median 80 μg/L (IQR 2–262).

< 10

10–49

50–99

100–199

HR (95% CI):

1.0 (Ref),

0.82 (0.59–1.14),

0.89 (0.58–1.36),

0.82 (0.64–1.05)

Adjusted for maternal socioeconomic status, education, BMI. Data also for As level > 200 μg/L
Viral infections serology

Cardenas, Smit, et al. (2015), USA/6–49 years,

cross‐sectional (NHANES 2003–2004, 2009–2010)

Seroprevalence of VZV IgG antibody

Seropositive/

seronegative

3250/98

u‐tiAs (1) geometric mean 6.57 μg/L (95% CI: 6.26, 7.9).

u‐As minus AsB / AsC (2) geometric mean (total As minus AsB and AsC) 5.64 μg/L (95% CI: 5.20, 6.12).

1‐unit increase in ln‐transformed u‐As (1) (μg/L)

1‐unit increase u‐As (2)

u‐As (1) GM (SE) seropositive versus seronegative: 6.77 (1.02) versus 8.31 (1.08), p: 0.01;

u‐As (2) GM (SE) seropositive versus seronegative: 5.85 (1.04) versus 7.62 (1.12), p: 0.02.

OR (95% CI):

1.87 (1.03, 3.44)

1.40 (1.00, 1.97)

Adjusted for age, sex, race, family poverty–income ratio, BMI classification, ln urinary creatinine, survey cycle.

Cardenas et al. (2016), USA/> 6 years,

cross‐sectional study (NHANES 2003–2012)

Hepatitis A antibodies (total anti‐HAV: IgG and IgM)

Seropositive/

seronegative 5064/6028

u‐tiAs (1) geometric mean 6.34 μg/L (SE 1.01).

u‐tAs minus AsB, AsC (2) geometric mean (total As minus AsB and AsC) 5.06 μg/L (SE 1.02)

1‐unit increase in ln‐transformed u‐As (1) (μg/L)

1‐unit increase u‐As (2)

OR (95% CI):

> = 2 doses vaccination 1·42 (1·11–1·81);

< 2 doses: 1·46 (0·83–2·59;

0 dose: 1·12 (0·98–1·30);

unknown vaccination history: 1·75 (1·22–2·52),

> = 2 doses vaccination: 1·17 (1·04–1·31)

< 2 doses: 1·26 (0·95–1·67)

0 dose: 1·07 (0·99–1·16)

unknown vaccination history: 1·20 (0·97–1·48)

Adjusted for log‐transformed creatinine, age, sex, race, family income/poverty ratio, country of birth, body mass index and survey year.

Cardenas et al. (2018), USA/> 6 years,

cross‐sectional study (NHANES 2003–2014)

Hepatitis B antibodies (total anti‐HAV: IgG and IgM)

Current HBV/

past HBV infection/

Vaccinated/

seronegative 33/636/3491/

8520

u‐As: DMA residual adjusted for arsenobetaine

For 1‐unit increase in ln‐transformed u‐As (DMA residual)

OR (95% CI):

for natural infection versus susceptible: 1.40 (1.15, 1.69);

natural infection versus vaccine induced immunity: 1.65 (1.34, 2.04);

seronegative versus vaccine induced immunity: 1.18 (1.05, 1.33)

Adjusted for log‐transformed creatinine, age, sex, race, family poverty‐income ratio, country of birth, BMI, survey year, log‐transformed serum cotinine, recent seafood consumption, self‐ reported Hepatitis B immunisation. Data also for how many doses of vaccination.

Hepatic effects
Islam et al. (2011), Bangladesh/15–60 years, cross‐sectional studyAST, ALT, ALPQuartiles Q1/Q2/Q3/Q4: 50/50/50/50

w‐As (μg/L):

Q1: 0.11–24.7,

Q2: 34–142,

Q3: 145–242,

Q4:249–546

hair As:

Q1: 0.05–1.43,

Q2: 1.57–2.80,

Q3: 2.81–5.52,

Q4: 5.66–37.24

nail As:

Q1: 0.15–3.17,

Q2: 3.21–6.25,

Q3: 6.26–11.27,

Q4: 11.4–37.42

Coefficient (95% CI) for w‐As and ALP:

1.0 (ref),

0.039 (−0.008, 0.085),

0.087 (0.041, 0.134),

0.165 (0.119, 0.212).

W‐As and AST:

1.0 (ref),

0.025 (−0.020, 0.070),

0.023 (−0.022, 0.067),

0.124 (0.080, 0.169).

w‐As and ALT:

1.0 (ref),

0.060 (−0.011, 0.132),

0.071 (0.00, 0.143)

0.186 (0.115, 0.257).

Hair As and ALP:

1.0 (ref),

0.028 (−0.023, 0.078),

0.044 (−0.005, 0.093),

0.117 (0.066, 0.168).

Hair As and AST:

1.0 (ref),

0.028 (−0.020, 0.075),

0.054 (0.008, 0.099),

0.096 (0.049, 0.144).

Hair As and ALT:

1.0 (ref),

0.064 (−0.010, 0.139),

0.086 (0.014, 0.158),

0.151 (0.076, 0.225).

Nail As and ALP:

1.0 (ref),

0.037 (−0.014, 0.088),

0.057 (0.005, 0.110),

0.109 (0.055, 0.163).

Nail As and AST:

1.0 (ref),

−0.007 (−0.055, 0.040),

0.010 (−0.039, 0.060),

0.062 (0.012, 0.113).

Nail As and ALT:

1.0 (ref),

0.043 (−0.033, 0.118),

0.055 (−0.024, 0.133),

0.084 (0.004, 0.165)

Adjusted for age, sex, BMI, smoking habit, skin lesions
Frediani et al. (2018), USA/> 12 years, cross‐sectional study (NHANES 2005–2014)ALT as a biomarker for NAFLDNAFLD yes/no: 2713/5805

u‐tAs minus AsB/AsC arsenocholine)

μg/L

Q1: < 2.42

Q2: 2.42–4.08

Q3: 4.08–6.99

Q4: > 6.99

1.00 (ref)

1.2 (0.7, 1.9),

1.6 (1.0, 2.8),

2.0 (1.2, 3.4)

Adjusted for age, survey cycle, gender, PIR, BMI Weight status, race/ethnicity.
Hearing loss
Li et al. (2018), Bangladesh/23–45 years, cross‐sectional studyHearing thresholdTotal 145, high versus low toenail As group: 97/48

Toenail As

High > 0.6 μg/g

Low < 0.6 μg/g

High versus low OR (95% CI):

1.28 (0.43–3.83) for 1 kHz,

4.27 (1.51–12.05) for 4 kHz,

3.91 (1.47–10.38) for 8 kHz,

4.15 (1.55–11.09) for 12 kHz

Adjusted for age, sex, smoking history, BMI.
Shokoohi et al. (2021), Iran/10–49 years, cross‐sectionalHearing threshold (Pure Tone Audiometry)Exposed/non‐exposed 120/120

w‐As (μg/L)

non‐exposed: 0.179

exposed: 200 (region A), 76.6 (region B), 74.5 (region C)

Exposed versus non‐exposed OR (95% CI):

2.03 (0.96–4.31) overall,

1.49 (0.24–9.18) for 2 kHz,

2.07 (0.92–4.62) for 4 kHz,

1.85 (0.74–4.66) for 8 kHz

Adjusted for age, smoking.
Allergy
Tsai, Wang, et al. (2021), Taiwan/4 years, birth cohort study (TMICS)Atopic dermatitis (questionnaire)Dermatitis yes/no: 110/260

u‐tiAs (inorganic) 30 trimester of pregnancy

2‐fold increase u‐As

Dermatitis versus control u‐As μg/g creatinine median (IQR):

30.91 (27.51–37.25) versus 28.62 (25.97–33.14), p < 0.001.

OR (95% CI):

2.42 (1.33–4.39)

Adjusted for child's sex, parental allergies, geographic area, exposure to tobacco smoke during pregnancy, exposure to tobacco smoke at age 4 years, maternal educational level. Data also for sensitivity analysis based on the existence of maternal allergy.

Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; As, arsenic; AsB, arsenobetaine; AsC, arsenocholine; AsMat, Health Consequences of Arsenic in Matlab; AST, aspartate transaminase; BMI, body mass index; CI, confidence interval; DMA, sum of dimethylarsinous acid and dimethylarsinic acid; GM, geometric mean; HAV, hepatitis A virus; HBV, hepatitis B virus; HEALS, Health Effects of Arsenic Longitudinal Study; HR, hazard ratio; IgG, immunoglobulin G; IgM, immunoglobulin M; IQR, interquartile range; IRR, incidence rate ratio; kHz, kilohertz; MINIMat, Maternal and Infant Nutrition Interventions in Matlab; MMA, sum of monomethylarsonous acid and monomethylarsonic acid; n, number; NAFLD, non‐alcoholic fatty liver disease; NHANES, National Health and Nutrition Examination Survey; NS, not significant; OR, odds ratio; PIR, poverty income ratio; Q, quartile; ref, reference; SE, standard error; tAs, total arsenic; TMICS, Taiwan Maternal and Infant Cohort Study; u‐As, urinary arsenic; u‐tiAs, urinary total inorganic arsenic (sum of iAs and its methylated metabolites MMA and DMA); USA, United States of America; VZV, Varicella‐Zoster Virus; w‐As, water arsenic.

-