Associations between Arsenic in Drinking Water and the Occurrence of End-Stage Renal Disease with Modifications by Comorbidities: A Nationwide Population-Based Cohort Study in Taiwan

Associations between Arsenic in

Drinking Water and the Occurrence of

End-Stage Renal Disease with Modifications

by Comorbidities: A Nationwide

Population-Based Cohort Study in Taiwan

Ya-Yun Cheng, How-Ran Guo

Department of Environmental and Occupational Health

College of Medicine, National Cheng Kung University

Tainan, Taiwan

27 Sep. 2017

Arsenic Important

Nevertheless, epidemiology studies on the

association between arsenic exposure and the

occurrence of ESRD are still limited.

60-75％

10-15％

10-20％

Blood

Liver

Liver

(Vahter M, 2000; Kitchin, 2001; Stýblo M et al., 2002; Hayakawa et al., 2005)

Monomethylarsonic Acid (MMA)

Dimethyarsinic Acid (DMA)

MMA3+＞As3+＞As5+＞MMA5+＝DMA5+

3

Arsenic concentrates in the kidney

during its urinary excretion that

affects the function of proximal tubules and glomerulus

Mechanisms pertaining to arsenic toxicity. Toxicol Int. 2011 Jul;18(2):87-93.

As ROSBUN+Cr

(Singh et al. 2011)

(Liao et al. 2009) ⁴

NCKU 5

Objective

1. To evaluate the associations between arsenic

exposure and the occurrence of ESRD in

Taiwan.

 We conducted a nationwide population-based

study including both the BFD endemic area

and other areas in Taiwan.

NCKU 6

Materials and methods

NCKU 7

Arsenic exposure index

 Data on well water arsenic were obtained from a

nationwide survey conducted by the Taiwan

Provincial Institute of Environmental Sanitation

using the standard mercuric bromide stain method.

(Lo et al.,1977; APHA, 1985)

 There were more than 80,000 wells, mostly

between 1974 and 1976, and were available for

311 townships (85%).

 According to the cut-offs adopted by the survey

reports, which was the regulatory standard at the

time of the survey, we defined a high arsenic level

as ≥ 50 µg/L

Jan 1,

1996

Jan 1,

1998

Dec 31,

2010

New ESRD cases once diagnosis

with Catastrophic illness code 001

Prevalent

CKD / ESRD

Follow up 13 years

Allowing at least 24 months wash-out period

LHID2000 Population-based ecological cohort study:

1,000,000 random sampling enrollees during 1996-2010

Adjust comorbidity

With 3 times OPD

diagnosis by a physician

SL Wang, 2003 National Health Insurance (NHI)

Longitudinal Health Insurance Database (LHID)

Evaluate the associations between arsenic exposure

and the occurrence of ESRD.

Drinking history

Address at 1998

Age ≥40 at 1998

NCKU 9

Risk factors of renal disease

 Epidemiological and clinical evidence have shown a

link between hypertension, diabetes, obesity, and

metabolic syndrome (Comorbidity) and the onset

and progression of CKD.

 SEX, Age, Edu, SES

 Western medicines



NSAID (non-steroidal anti-inflammatory drugs)



Acetaminophen

 Chinese herbal medicine (Aristolochic acid)

(Vanherweghem JL, 1993; Yang CS, 2000; Chang CH, 2001; Yang HY, 2006)

Disease ICD-9-CM / A-CODE Criteria

ESRD Dialysis 585-586 combined with copayment code ‘‘001’’

(indicated the presence of a Catastrophic illness)

1 OPD visits

Hypertension 401- 405, A260 ≧3 OPD visits

Diabetes 250, A181 ≧3 OPD visits

Hyperlipidemia 272.0-272.4, A182 ≧3 OPD visits

CAD 410-414 ≧3 OPD visits

CHF 428.0 ≧3 OPD visits

Stroke 430-438 ≧3 OPD visits

Anemia 280-285 ≧3 OPD visits

Assessment of renal disease and comorbidity

International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM)

NCKU 11

Statistic

• Chi-square test, T-test, ANOVA

• Cox PH regression (Survival analysis): ESRD

 Kaplan–Meier, Log-rank

 Cox regression: Hazard Ratio, 95% C.I.

 Single, multiple, backward stepwise (inclusion set at

p<0.05; exclusion set at p>0.15)

• SAS 9.3 + SPSS1 17.0

• Two-side p<0.05

NCKU 12

123 Townships with high

arsenic levels (≥ 50 µg/L)

were clustered in the

southwestern and

northeastern coast regions.

Fig. 1 Map of cumulative

incidence rate of ESRD

with age ≥ 40 years and

arsenic from drinking water

on each township in Taiwan

13

Fig. 2 Flow chart of the

nationwide cohort study

1.4%

Arsenic data covered

93.5% population

1.7%

NCKU 14

1-1

TABLE-1 Cox regression models for occurrence of end-stage renal disease

(N=362,505).

NCKU 15

1-2

TABLE-1 Cox regression models for occurrence of end-stage renal disease

(N=362,505).

NCKU 16

1-3

TABLE-1 Cox regression models for occurrence of end-stage renal disease

(N=362,505).

1~2 3~7

NCKU 17

TABLE-2 Cox models for the effect modification of comorbidity on 2-1

the association between arsenic exposure and end-stage renal disease.

  

  

NCKU 18

Fig. 3 Survival analysis of Kaplan-Meier curve

with end-stage renal disease (ESRD) by

number of comorbidities.

 Log-Rank test: p < 0.05 

Strengths and limitations

 A large sample size and a long duration: therefore,

the results should be more reliable than those from

smaller case-control or cross-sectional studies.

 Ecological fallacy.

 Ours study using resident area measurements

instead of personal exposure data: series of studies

have been conducted using similar approaches and

identified increased risk of cerebrovascular

disease, pterygium, and liver cancer.

(Lin et al., 2008; Cheng et al., 2010; Chung et al., 2013; Lin et al., 2013)

NCKU 19

Conclusion

 Using 50 µg/L as the cut-off, we found that a higher

level of arsenic in the drinking water was positive

associated with AHR of 1.12 (95% CI: 1.05-1.18, p <

0.001) for the occurrence of ESRD in Taiwan,

independent of most documented risk factors including

age, sex, SES, Hypertension, Hyperlipidemia, Stroke,

CAD, CHF, DM, and Anemia.

 We observed a effect modification of arsenic on ESRD

by comorbidities, especially in less than two.

 Intervention programs for the prevention and control of

comorbidities should be implemented in endemic areas

of exposure to slow the occurrence of renal disease

especially on early CKD.

NCKU 20

Acknowledgements

 This work was supported by Grants NSC102-

2314-B-006 -026-MY2 and MOST-104-2314-B-

006 -021 -MY2 from the Ministry of Science and

Technology of Taiwan, R.O.C.

 The database in our study was established by the

National Health Insurance Research Database

(NHIRD).

 Co-author: How-Ran Guo (M.D., Sc.D.), Junne-Ming Sung

(M.D.), Yu-Tzu Chang (M.D.)

 Travel Awards from National Cheng Kung

University (NCKU) and Foundation for the

Advancement of Outstanding Scholarship (FAOS).

NCKU 21

NCKU 22

Thank you for

your attention！

Q & A

NCKU 23

NCKU 24

2014 Environmental Health Perspectives • volume 122 P.213 2009 The Lancet Oncology, Volume 10, Issue 5, Pages 453 - 454

http://research.stevens.edu/index.php/sers-on-site-analysis

This graphic illustrates global occurrences of arsenic in groundwater

In TAIWAN 0.5 million

were exposed to arsenic concentration

in drinking water above 50 µg/L

which was the regulatory standard at the time of survey.

(Lo et al. 1977, Guo et al. 1997; Guo et al. 1998)

As in well water: 350 to 1140 µg/L

1910s ~ the early 1970s

Arsenic-exposed areas in Taiwan

Fig. Map of southwestern and northeastern arsenic-exposed areas in Taiwan. There

are four townships in the southwestern exposed area: Putai (1), Ichu (2), Peimen

(3), and Hsuehchia (4); and four townships in the northeastern exposed area:

Chiaohsi (1), Chuangwei (2), Wuchieh (3), and Tungshan (4).

(Guo et al.,1998, 2008) 26

As in well water: 0.15μg/L ~3590μg/L

The late 1940s to the early 1990s

NCKU 27

Mechanisms pertaining to arsenic toxicity. Toxicol Int. 2011 Jul;18(2):87-93.

Arsenic-Induced Nephro-toxicity

By ROS (Reactive Oxygen Species)

Enhances lipid peroxidation and

cellular damage in renal tissue.

Acute renal dysfunction due to

arsenic exposure is characterized

by acute tubular necrosis and

cast formation with increasing in

blood urea nitrogen (BUN) and

creatinine (Cr) levels.

(Giberson et al. 1976, Sasaki et al. 2007)

27

Epidemiology: As – kidney disease

Reference Design Arsenic exposure Main findings

(Chen et al. 2011) Chemosphere

Community-based cross-sectional study from central Taiwan

Urine As U-As might relate to renal

dysfunction even other important risk factors were taken into

account. (β2MG > 0.154 mg/L) U-As > 35 lg/g

(eGFR < 90 mL/min/1.73 m2/year) U-As > 75 lg/g (Hsueh et al. 2009)

American Journal of Kidney Diseases

Hospital based Case- control study

125 CKD patients and 229 controls

Urine As Total arsenic level was

associated significantly with CKD in a dose-response relationship.

(Chiu and Yang 2005) Journal of Toxicology and Environmental Health - Part A

Standardized mortality ratios

(SMRs) for the years 1971–2000.

Residents BFD endemic area in the southwestern coast of Taiwan more than 50 yr.

(SMRs) for renal diseases were positively correlated.

(Chiou et al. 2005a) International Journal of Epidemiology

National Health Insurance (NHI) database.

1999-2000 prevalence

Residents BFD endemic area in the southwestern coast of Taiwan

An increased prevalence of microvascular diseases,

including neurological and renal disorders, is associated with arsenic ingestion.

NCKU 28

Arsenic level was positively associated significantly

with kidney disease and renal dysfunction.

28

NCKU 29

American Journal of Kidney Diseases, 2007 Vol 49, No 1 : pp 46-55

Epidemiology of

Chronic Kidney Disease

In NHIRD

(Kuo et al. 2007) (Wen et al. 2008)

They were significantly higher

in southwestern Taiwan with

an increasing trend.

NCKU 30

（United States Renal Data System, USRDS 2012）

12% CKD 2.7 million in 2012

0.26% ESRD cost NHI: 7 % $

30